Plan performance — Medical loss ratio

Anchor Health Cooperative

Plan performance

A synthetic demo surface — "Anchor Health Cooperative" is a fictional, obviously-fake HMO. No real payer, no PHI. Every number here reconciles to the vault: premium = Σ fct_premium_month, plan-paid = Σ fct_claim (status = paid). The app carries no business logic — the loss ratio is computed in the fct_loss_ratio gold mart.

Medical loss ratio · 2017–present

The single most legible "is this plan healthy?" number: plan-paid claims ÷ premium revenue. A typical group plan runs 0.80–0.90 — most of every premium dollar pays for care, the rest covers administration and margin.

Blended MLR (2017+)

85.3%

Premium billed (money in)

$1,279,353

Claims paid (money out)

$1,090,657

For every $1 of premium, about $0.85 pays for care — a healthy group loss ratio.

By plan segment

Each claim is attributed to the segment of the member's coverage in the claim's service month (not a dominant segment) — a claim belongs to the plan that was active when it happened. Every segment lands in the healthy band, bracketing the blended number.

No Results

MLR trend by year

The blended loss ratio year over year — a line that stays inside the 0.80–0.90 band as premium and claims track together across the aligned window.

Reconciles to the vault (check it live): premium = Σ fct_premium_month.billed_premium_amt; plan-paid = Σ fct_claim.total_paid_amt where claim_status = 'paid'. Orphan (never-covered member) plan-paid = $0.

Risk-adjusted loss ratio · the sophisticated sequel

Every segment's raw MLR sits inside the healthy 0.80–0.90 band — but that already reflects risk, because premium is risk-graded: MEDICARE members are billed ~$29/month and COMMERCIAL ~$16, precisely because MEDICARE members are sicker. The new signal is the relative risk — how sick each segment's members are versus the whole scored population, from the HCC-style member risk score (cs_member_risk_score). Read the two numbers together: a segment running N× sicker than average while holding an MLR of ~0.85 is managing that risk efficiently; a healthier segment at the same MLR less so.

Population avg risk score

0.59

DUAL relative risk (× population)

2.32

MEDICARE relative risk (× population)

1.80

Each segment's risk score is averaged over its claim population — the members whose paid claims land in that segment by service month, the same attribution the raw MLR uses to split paid dollars. The chart pairs the two headline numbers side by side.

No Results

Read the two numbers together: DUAL (~2.3× sicker) and MEDICARE (~1.8× sicker) each hold an MLR near 0.85 despite carrying by far the heaviest clinical burden — that is the efficiency story the raw MLR alone can't tell. COMMERCIAL and MEDICAID run below average risk, so their similar ~0.83–0.87 MLRs are less impressive once you know how light their risk is. The efficiency index (raw MLR ÷ relative risk) compresses that into one number, but note it is not a literal MLR: premium is already risk-graded, so dividing nets out risk a second time and the index runs out-of-band for the sickest segments — read it as a relative signal, not a loss ratio. The score lives on the clinical patient (cs_member_risk_score, keyed on hk_patient) and bridges to the payer member through cs_patient_member_same_as — the clinical-and-claims join on the resolved person that is the platform's core thesis.