The Evidence Behind HEDIS COL: Colorectal Cancer Screening

Key findings

MY2026 specification, age band, modalities. NCQA’s COL-E measures the percentage of members 45–75 with one of: colonoscopy in the measurement year (MY) or 9 prior years; CT colonography in MY or 4 prior years; flexible sigmoidoscopy in MY or 4 prior years; multi-target stool DNA (FIT-DNA/Cologuard) in MY or 2 prior years; or FIT/gFOBT during the MY. The age expansion from 50–75 to 45–75 took effect for MY 2022 following the USPSTF May 2021 update.

ECDS-only transition. NCQA confirms that COL is among the first four hybrid HEDIS measures already transitioning to ECDS (alongside CCS, CIS, IMA), with parallel hybrid/ECDS comparison reporting permitted in MY 2024 and MY 2025 and ECDS as the standard pathway for MY 2026 forward (NCQA, “NCQA’s Proposed Timeline for Retiring and Replacing HEDIS Hybrid Measures,” Nov 15, 2024).

Stars status. COL is included in 2026 Medicare Advantage Star Ratings as a Process Measure with weight = 1 (CMS, “2026 Star Ratings Measures and Weights”). Per CMS’s “2027 Star Ratings Measures and Weights,” the respecified COL-E is treated as a “new measure” for 2027, which under 42 CFR §423.186(e)(2) again receives weight 1 for its first year.

USPSTF: Grade A for 50–75, Grade B for 45–49. The May 18, 2021 USPSTF final recommendation (Davidson et al., JAMA 2021;325:1965-1977) extended the screening age floor from 50 to 45; the Medicare Physician Fee Schedule CY2023 final rule (87 FR 69404, effective Jan 1, 2023) aligned Medicare coverage with that age floor and eliminated cost-sharing for follow-on colonoscopy after a positive non-invasive stool test.

Mortality reductions vary by modality. Strongest RCT evidence: ~33% CRC mortality reduction with annual rehydrated gFOBT in the Minnesota Colon Cancer Control Study at 13 years (Mandel 1993, NEJM); a pooled 16% relative CRC mortality reduction (RR 0.84, 95% CI 0.78–0.90) across four RCTs and ~320,000 participants in the Cochrane review (Hewitson et al., Am J Gastroenterol 2008); 26% reduction in CRC incidence and 30% in CRC mortality at 17 years in the UK Flexible Sigmoidoscopy Screening Trial (Atkin 2017, Lancet); and an 18% incidence/10% mortality reduction at 10 years in the intention-to-screen arm of NordICC (Bretthauer 2022, NEJM), with per-protocol estimates of approximately 31% incidence and 50% mortality reduction in those actually screened.

The follow-up problem. In the Kaiser Permanente cohort of 70,124 patients with abnormal FIT, Corley et al. (JAMA 2017;317:1631-1641) found that colonoscopy delayed more than 10 months after the positive FIT was associated with significantly higher CRC incidence and more advanced-stage disease at diagnosis; in the VA cohort of 204,733 patients (San Miguel/May et al., Gastroenterology 2021;160:1997-2005), CRC-specific mortality was 52% higher (HR 1.52, 95% CI 1.15–1.99) for follow-up at 19–21 months versus 1–3 months. This is the central reason NCQA is developing a dedicated follow-up-after-abnormal-CRC- screening HEDIS measure.

Disparities are large and persistent. Per BRFSS 2022 data summarized by CDC (Preventing Chronic Disease 2024), 61.7% of Hispanic adults were up-to-date with CRC screening versus 74.6% non-Hispanic White and 75.3% non-Hispanic Black. Per NCQA (“NCQA Is Developing a New HEDIS Measure for Colorectal Cancer Screening Follow-Up”), MY2023 plan-level rates were 56% for commercial, 64% for Medicare, and 39% for Medicaid - a 25-point Medicaid-Medicare gap. Rural Medicaid members and recent immigrants have markedly lower screening rates.

What works to close the gap. Mailed FIT outreach produced a pooled absolute increase of 28% across all stool-test outreach studies (27% for FIT-specific, 28% for gFOBT-specific) in 7 RCTs and 12,501 patients (Jager M, Demb J, Asghar A, et al., Dig Dis Sci 2019;64(9):2489-2496). Patient navigation increases CRC screening 64% relative to usual care (pooled RR 1.64, 95% CI 1.42–1.92, 22 RCTs; Ali-Faisal et al., J Gen Intern Med 2020). Combined mailed FIT plus navigation produces the largest gains in safety-net settings.

1) The HEDIS COL-E measure (MY 2026)

Full name. Colorectal Cancer Screening (HEDIS abbreviation: COL-E for the ECDS-reported version; legacy hybrid abbreviation: COL). The measure is part of the Effectiveness of Care domain.

Eligible population (denominator). Members aged 45–75 as of December 31 of the measurement year, with continuous enrollment for the measurement year and the year prior (with an allowable gap of one 45-day period or less per year). Required for Commercial, Medicaid, and Medicare product lines.

Numerator (compliant screening). One or more of the following, with documented date and result:

• Colonoscopy during the MY or any of the 9 years prior (effective 10-year lookback).
• CT colonography during the MY or any of the 4 years prior (5-year lookback).
• Flexible sigmoidoscopy during the MY or any of the 4 years prior (5-year lookback).
• FIT-DNA / multi-target stool DNA test (Cologuard) during the MY or any of the 2 years prior (3-year lookback).
• FIT or gFOBT during the MY (annual). FOBT performed in-office or on a digital-rectal-exam specimen does not count.

Required exclusions.

• History of colorectal cancer at any time (ICD-10 C18.0–C18.9, C19, C20, C21.2, C21.8, C78.5, Z85.038, Z85.048).
• History of total colectomy (CPT 44150–44153, 44155–44158, 44210–44212; ICD-10-PCS 0DTE0ZZ, 0DTE4ZZ, 0DTE7ZZ, 0DTE8ZZ).
• Members in hospice or who use a hospice benefit during the MY.
• Members who received palliative care during the MY (e.g., Z51.5).
• Members who died during the MY.
• Members 66+ enrolled in an Institutional SNP or in long-term institutional (LTI) care any time in the MY.
• Members 66+ with both frailty (≥2 frailty indicators on different dates of service in the MY) and advanced illness (≥2 outpatient/ED/non-acute claims on different dates with an advanced-illness diagnosis OR ≥1 acute inpatient claim with an advanced-illness diagnosis OR a dispensed dementia medication, in the MY or prior year).

Notably, for MY 2025, NCQA converted COL’s previously optional exclusions for advanced illness/frailty and LTI to required exclusions, removing high-frailty older adults from the denominator and raising achievable rates.

Reporting structure. Per the NCQA blog “NCQA’s Proposed Timeline for Retiring and Replacing HEDIS Hybrid Measures” (Nov 15, 2024), “Four hybrid measures-Colorectal Cancer Screening, Cervical Cancer Screening, Childhood Immunization Status and Immunizations Status for Adolescents-are already transitioning to ECDS. These measures have parallel reporting options, allowing plans to compare ECDS results to traditional hybrid results for MY 2024 and MY 2025.” From MY 2026, ECDS becomes the standard pathway. This eliminates the annual hybrid chart-chase; any compliance signal from a chart must arrive as prospective supplemental data (claims, EHR feeds via FHIR, registries, HIE).

Medicare Star Ratings. COL is a CMS Star Ratings measure for Medicare Advantage. Per the CMS “2026 Star Ratings Measures and Weights,” Colorectal Cancer Screening is classified as a Process Measure with weight = 1 in Part C, contrasted with intermediate-outcome measures (e.g., Controlling Blood Pressure, Diabetes Care - Blood Sugar Controlled) and outcome measures (e.g., Plan All-Cause Readmissions) that carry weight 3. The CMS 2025 Star Ratings fact sheet noted that “an increase in scores for contracts at the lower end of the distribution for some measures, such as Colorectal Cancer Screening (Part C) pushed cut points higher for those measures.” Per the “2027 Star Ratings Measures and Weights,” “The Colorectal Cancer Screening measure is also being replaced by a respecified version and is treated as a new measure,” restarting it at weight 1 for 2027 per 42 CFR §423.186(e)(2).

2) Code sets (representative - not exhaustive)

Plans should consume the NCQA value sets directly, but the high-yield codes for numerator capture include:

• Colonoscopy - CPT 44388–44394, 44397, 44401–44408, 45355, 45378–45393, 45398; HCPCS G0105, G0121.
• Flexible sigmoidoscopy - CPT 45330–45335, 45337, 45338, 45340–45342, 45346, 45347, 45349, 45350; HCPCS G0104.
• CT colonography - CPT 74261, 74262, 74263; LOINC 60515-4, 72531-7, 79069-1, 79071-7, 79101-2, 82688-3.
• FIT-DNA (Cologuard) - CPT 81528; HCPCS G0464 (legacy); LOINC 77353-1, 77354-9.
• FIT/gFOBT - CPT 82270, 82274; HCPCS G0328 (immunoassay); LOINC 12503-9, 12504-7, 14563-1, 14564-9, 14565-6, 2335-8, 27396-1, 27401-9, 27925-7, 27926-5, 29771-3, 56490-6, 56491-4, 57905-2, 58453-2, 80372-6.
• Exclusion codes - ICD-10 C18–C20, Z85.038, Z85.048 (CRC history); CPT 44150–44158, 44210–44212 and ICD-10-PCS 0DTE-series (total colectomy); Z51.5 (palliative care).

Modifier 33 (preventive) and modifier PT (screening converted to diagnostic/therapeutic) are critical for correct ACA/Medicare cost-share processing.

3) Clinical evidence base

USPSTF 2021 (Davidson KW et al., JAMA 2021;325(19): 1965-1977; doi:10.1001/jama.2021.6238). The Task Force gave Grade B for 45–49 and Grade A for 50–75 (selective screening 76–85, Grade C). All major modalities - colonoscopy, flexible sigmoidoscopy ± FIT, CT colonography, FIT, gFOBT, and stool DNA - are endorsed without ranking one above another. ACA §2713 obligates non-grandfathered private plans (and Medicaid expansion plans) to cover Grade A/B preventive services without cost-sharing.

Modality-specific effect sizes

• Annual gFOBT (Minnesota Colon Cancer Control Study). In 46,551 randomized adults aged 50–80, Mandel et al. (NEJM 1993;328:1365-1371) reported a 33% relative reduction in CRC mortality with annual rehydrated guaiac FOBT at 13 years; Mandel et al. (NEJM 2000;343:1603-1607) extended this to demonstrate a 20% reduction in CRC incidence at 18 years through polyp detection and removal.
• Pooled FOBT (Cochrane). Hewitson et al. (Am J Gastroenterol 2008;103:1541-1549) found a 16% pooled RR reduction in CRC mortality (RR 0.84, 95% CI 0.78–0.90) across four RCTs (~320,000 participants), rising to 25% (RR 0.75, 95% CI 0.66–0.84) for attenders.
• Flexible sigmoidoscopy. Atkin et al. (Lancet 2017;389:1299-1311; 17-year follow-up of the UK Flexible Sigmoidoscopy Screening Trial, 170,432 adults) reported a 26% reduction in CRC incidence and 30% reduction in CRC mortality in intention-to-screen analysis. Atkin/Wooldrage et al. (Lancet Gastroenterol Hepatol 2024; 21-year follow-up) confirms durable effects out to two decades from a single screen. Holme et al. (JAMA 2014;312:606-615) and Schoen et al. (NEJM 2012;366:2345-2357) provide corroborating Norwegian and U.S. PLCO evidence.
• Colonoscopy. NordICC (Bretthauer M et al., NEJM 2022;387:1547-1556; 84,585 adults aged 55–64 in Norway, Poland, Sweden) found at 10 years an 18% reduction in CRC incidence (RR 0.82, 95% CI 0.70–0.93) and a non-significant 10% reduction in CRC mortality (RR 0.90, 95% CI 0.64–1.16) in the intention-to-screen population - driven down by 42% screening uptake. Per-protocol (adjusted) estimates were approximately 31% incidence and 50% mortality reduction in those who actually underwent colonoscopy. The 2026 DDW update reported a sustained ~30% incidence reduction without significant mortality reduction at extended follow-up.
• Multi-target stool DNA (Cologuard). Imperiale TF et al. (NEJM 2014;370:1287-1297; 9,989 average-risk adults) reported single-test sensitivity for CRC of 92.3% with mt-sDNA vs. 73.8% with FIT, and for advanced precancerous lesions 42.4% vs. 23.8% - at the cost of lower specificity (86.6% vs. 94.9%). The next-generation Cologuard Plus (BLUE-C study; Imperiale et al., NEJM 2024;390:984-993) reported CRC sensitivity of 93.9% and advanced precancerous lesion sensitivity of 43.4%.

Lives saved (population-level)

USPSTF/CISNET microsimulation (Knudsen et al., JAMA 2021;325:1998-2011) estimated that screening 1,000 average-risk adults from age 45 to 75 with colonoscopy every 10 years yields ~24 life-years gained and ~24 CRC cases averted versus no screening, and starting at 45 vs. 50 yields an incremental 22–27 life-years per 1,000 across modalities.

Per Siegel RL, Wagle NS, Star J, Kratzer TB, Smith RA, Jemal A (“Colorectal cancer statistics, 2026.” CA Cancer J Clin. 2026;76:e70067), “There will be an estimated 158,850 new cases of CRC in the United States in 2026” (108,860 colon, 49,990 rectal) and 55,230 deaths, with “incidence rates increased by 3% annually in adults aged 20–49 years” during 2013–2022.

Follow-up after abnormal stool test (the most important operational evidence)

• Corley DA, Jensen CD, Quinn VP, et al., JAMA 2017;317:1631-1641 (Kaiser Permanente Northern California, 70,124 patients with abnormal FIT): compared with follow-up colonoscopy at 8–30 days, follow-up at 10–12 months was associated with adjusted OR 1.48 for any CRC (95% CI 1.05–2.08), with risk of advanced-stage CRC rising steeply beyond 10 months.
• San Miguel Y, Demb J, Martinez ME, Gupta S, May FP, Gastroenterology 2021;160:1997-2005 (VA, 204,733 patients with abnormal FOBT/FIT): CRC mortality HR 1.52 (95% CI 1.15–1.99) for follow-up at 19–21 months versus 1–3 months.
• Meester RG, Zauber AG, Doubeni CA, et al., Clin Gastroenterol Hepatol 2016;14:1445-1451: modeled a 16% increase in CRC mortality with one-year delay vs. 2-week follow-up.

These data underpin NCQA’s active development of a new HEDIS measure for follow-up after abnormal CRC screening, funded by the CDC.

4) Disparities and equity

Disparities cut along race/ethnicity, payer, and geography:

• Race/ethnicity. Per BRFSS 2022 (CDC, Preventing Chronic Disease 2024), up-to-date CRC screening was 75.3% in non-Hispanic Black, 74.6% in non-Hispanic White, 61.7% in Hispanic, and approximately 60–61% in non-Hispanic Asian American adults. American Indian/Alaska Native populations are the only racial/ethnic group whose CRC mortality is not declining (Siegel et al., CA Cancer J Clin 2023).
• Payer (NCQA MY 2023 plan reporting). 56% commercial, 64% Medicare, 39% Medicaid - a 25-point Medicaid-Medicare gap (NCQA, “NCQA Is Developing a New HEDIS Measure for Colorectal Cancer Screening Follow-Up”).
• Geography. Rural residents on Medicaid have particularly low screening - Coronado et al. (JAMA Netw Open 2025, cluster RCT in 28 rural Oregon Medicaid clinics, n=5,614) reported a 7.3 percentage-point absolute increase via a mailed FIT + navigation intervention over usual care.
• Equity-relevant policy levers. The 2022 CMS rule eliminating beneficiary cost-sharing for follow-on colonoscopy after a positive stool test (CY2023 PFS Final Rule, 87 FR 69404) and the parallel ACA FAQ Part 51 guidance (DOL/HHS/Treasury, Jan 10, 2022) removed an important financial barrier that disproportionately affected low-income and minority patients.

5) Operational/documentation failure modes

Even when screening is clinically done, the HEDIS gate often fails. The common failure modes in EHRs:

• Outside colonoscopies as unstructured “history.” A patient’s outside colonoscopy from 7 years ago lives as free text in an H&P note, a scanned PDF, or a problem-list comment - none of which an administrative-only HEDIS pull can ingest. Under hybrid reporting, plans recovered some via chart-chase; under MY 2026 ECDS-only COL, these must arrive as structured supplemental data (LOINC-coded result, CPT code, or a discrete “history of colonoscopy + date” element).
• FIT/FOBT documentation gaps. Many EHRs file FIT results as a normal-range lab without a HEDIS-reportable LOINC; some practices distribute FIT kits without an associated order, so the kit’s return generates a result with no order trail. In-office gFOBT and DRE-based FOBT are explicitly excluded yet are still billed.
• Cologuard ordered outside the plan network. Exact Sciences ships and processes most kits centrally; if the ordering provider is non-network, the result may never flow through plan claims.
• Exclusion documentation. Total colectomy is frequently captured only in an operative report from years past, not as a current diagnosis or procedure code on a recent claim. Hospice and palliative-care indicators require an active claim (Z51.5 or a hospice enrollment code) - informal “comfort care” notes don’t qualify.
• Advanced illness + frailty. Both must be present; documentation often captures one but not the other (e.g., dementia medication dispensed without a frailty indicator like falls, walker use, weakness).

Where ECDS reporting changes the game. Under ECDS, the eligible population is the full member panel (no 411-member sample) and the only way to credit screening done outside the plan’s claims pipeline is through structured supplemental data - FHIR feeds from EHRs, HIE pulls, immunization/registry-style clinical repositories, and case-management systems. This makes interoperability and AI-assisted extraction (NLP from notes, OCR from scanned outside reports, LOINC mapping) operationally decisive rather than nice-to-have.

6) Quality improvement evidence

• Mailed FIT outreach. Jager M, Demb J, Asghar A, et al. (Dig Dis Sci 2019;64(9):2489-2496) pooled 7 RCTs and 12,501 patients to find a pooled absolute increase of 28% across all stool-test outreach studies (27% for FIT-specific, 28% for gFOBT-specific). STOP CRC (Coronado et al., JAMA Intern Med 2018;178:1174-1181) demonstrated EHR-embedded mailed FIT outreach significantly increased FIT completion in federally qualified health centers.
• Patient navigation. Ali-Faisal et al. (J Gen Intern Med 2020;35:3026-3035) pooled 22 RCTs and found a 64% relative increase in CRC screening (RR 1.64, 95% CI 1.42–1.92). The Community Preventive Services Task Force recommends patient navigation for CRC screening on the basis of a meta-analysis (RR 1.65, 95% CI 1.38–1.99 across 12 studies) and finds it cost-effective.
• Combined mailed FIT + navigation. Coronado et al. (JAMA Netw Open 2025) cluster-RCT in 28 rural Oregon Medicaid clinics (n=5,614) showed a 7.3-percentage-point absolute increase in CRC screening over usual care.
• EHR reminders and point-of-care prompts. Dougherty et al. (JAMA Intern Med 2018;178:1645-1658; 73 RCTs) found that “outreach methods” (mailed FIT, mailed reminders) and patient navigation produced the largest and most consistent gains; clinician-only reminders had smaller and less consistent effects.
• Text messaging and behavioral nudges. Mehta et al. (Am J Gastroenterol 2018;113:1848-1854; J Gen Intern Med 2021;36:2858-2864) demonstrated that opt-out framing and serial text reminders meaningfully increase FIT return in urban safety-net populations.

7) Regulatory & policy context

• USPSTF 2021 (Davidson KW et al., JAMA 2021;325(19):1965-1977). Lowered floor to 45 (Grade B for 45–49, Grade A for 50–75).
• ACA §2713. All non-grandfathered private plans must cover USPSTF Grade A/B preventive services without cost-sharing.
• ACA FAQs Part 51 (DOL/HHS/Treasury, January 10, 2022). Clarified that colonoscopy after a positive non-invasive screening test is part of the screening continuum and must be covered without cost-sharing for plan years beginning on or after May 31, 2022.
• Medicare PFS CY2023 Final Rule (87 FR 69404, published Nov 18, 2022; effective Jan 1, 2023). Reduced the Medicare CRC-screening age floor from 50 to 45 and reclassified the follow-on colonoscopy after a positive non-invasive test as a “complete screening” rather than a diagnostic procedure, eliminating coinsurance/deductible for the follow-up colonoscopy.
• CAA 2021 phase-down. For 2022, beneficiary coinsurance on a screening colonoscopy that finds a polyp dropped to 20%; for 2023–2026 it is 15%; from 2027 forward it is 0%.

8) Star Ratings implications for MA plans

COL is included in 2026 Star Ratings as a Part C Process Measure with weight 1 (CMS, “2026 Star Ratings Measures and Weights”). In the 2027 Star Ratings, the respecified COL-E is treated as a new measure and again receives weight 1 for its first year, per 42 CFR §423.186(e)(2). Although a single weight-1 measure is small in isolation, COL frequently sits at or just below the 4-star cut-point line for many contracts, and the 2025 Star Ratings fact sheet specifically called out upward cut-point movement on COL. Combined with Tukey outlier deletion (effective 2024 Star Ratings) and the Health Equity Index reward (replacing the reward factor in 2027 Star Ratings), small absolute movements in COL performance compound across the rating.

Recommendations

For a health plan or risk-bearing provider organization seeking to improve COL performance:

1. Stand up an ECDS-ready screening registry now. By MY 2026, chart-chase is no longer a fallback. Move all numerator-positive events (CPT, HCPCS, LOINC) and exclusion events (total colectomy, hospice, palliative, advanced illness + frailty) into discrete, FHIR-queryable fields. Benchmark: ≥95% of known outside colonoscopies should be ingestible as structured data within 90 days of receipt.
2. Default to mailed FIT for unscreened average-risk members aged 45–75. This is the single highest-yield, lowest-cost intervention with the strongest RCT evidence - a 28-percentage-point pooled absolute increase in 7 RCTs (Jager 2019). Benchmark: in a Medicaid population starting under 40% baseline, an 8–15 percentage-point absolute uplift over 12 months is realistic.
3. Treat any positive FIT or positive Cologuard as a sentinel event. Build a closed-loop tracking system that flags time-to-follow-up colonoscopy; the goal should be median ≤30 days, with no patient passing 6 months unfollowed. The Corley (JAMA 2017) and San Miguel/May (Gastroenterology 2021) data justify treating delays >6 months as a clinical safety event.
4. Layer patient navigation on the highest-risk segments. Concentrate human navigators on rural Medicaid, Hispanic and limited-English-proficiency populations, and any member with an abnormal stool test. Evidence supports a 50–65% relative uplift (Ali-Faisal 2020; Community Preventive Services Task Force).
5. Code exclusions accurately. Many plans understate the true denominator-eligible population because total colectomy, hospice, palliative care, and advanced-illness+frailty patients are not coded with the right combination of claims. The 2025 NCQA shift of these from optional to required exclusions makes accurate exclusion coding rate-positive.
6. Use AI for two distinct workstreams. (a) NLP/OCR to extract outside colonoscopy and stool-test results from scanned outside records, faxes, and unstructured notes, then write them back as ECDS-compliant supplemental data; (b) risk-stratified, behaviorally informed outreach (opt-out framing, serial reminders, language-concordant outreach) for unscreened members.

Thresholds that should change the recommendations. If a contract’s COL rate is already ≥75% and the follow-up-after-abnormal-FIT rate is ≥80% within 6 months, marginal investment in additional outreach yields little; resources should pivot to the upcoming NCQA follow-up-after-abnormal-CRC-screening measure and to Health Equity Index–relevant subpopulations. If baseline is <50%, mailed FIT outreach is the clear first move; navigation and EHR optimization are second-order.

Caveats

• The NordICC mortality finding (10-year intention-to-screen analysis) is statistically non-significant; the per-protocol estimate is consistent with prior observational data but should not be cited as a definitive RCT mortality reduction. The 2026 DDW update similarly did not show significant mortality reduction at extended follow-up. Citing colonoscopy as “proven to reduce mortality in an RCT” overstates current evidence; the more defensible claim is that colonoscopy reduces CRC incidence in an RCT and reduces mortality in cohort studies and in per-protocol analyses.
• Cologuard’s improved sensitivity comes with lower specificity than FIT, meaning more false positives and downstream colonoscopies; cost-effectiveness analyses are sensitive to assumed adherence patterns.
• The exact MY-by-MY transition status of COL hybrid reporting has been described inconsistently across NCQA documents; some third-party sources call MY 2024 the “last hybrid year,” while NCQA’s own blog frames MY 2024 and MY 2025 as a parallel-reporting window with ECDS as the standard pathway from MY 2026 forward. Plans should consult the current NCQA HEDIS Volume 2 Technical Specifications for definitive logic.
• Star Ratings cut points for COL are published annually by CMS in the Part C & D Star Ratings Technical Notes and shift each year; the exact 4-star threshold should be pulled from Table K-3 of the current Technical Notes rather than relying on aggregate commentary.
• Screening rate disparities by race/ethnicity narrowed in some surveys after the 2021 USPSTF age expansion but remain large for Hispanic and Asian American populations; aggregate national figures obscure substantial state-by-state and plan-by-plan variation.