PLAN Explore › Segmentation ◇ HTN · Chicagoland · 2024

Patient Segmentation

Slice the Market into segments and compare them on size, clinical depth, and value-based overlap.
Plan mode — sizing & opportunity. Which segments are worth pursuing?
This is a category, not a screen. Every segmentation insight is the same machine: pick a segment axis — the trait you group patients by, shown as the rows — and compare those groups on a shared metric set (the columns) plus a condition-specific clinical-trigger panel. Care model is the lead cut; switch the axis above the table to watch the same layout hold. Synthetic Chicago-area hypertension data.
A vendor sizing this market needs to know who is already under care management and by whom, how clinically addressable each slice is, and how much is locked up by value-based entities. This is descriptive Segmentation — the analysis you study before you commit to a Target list.
Grow mode. Same segments, same market math — your Panel is now overlaid on every row: how many of your patients fall in each segment, your penetration of that segment, and where your adherence trails the market. You are just another series on the same grid.
Segment by
Condition
Segment composition
Segmentation matrix
1 row per segment · cells < 11 benes suppressed
The hero. Each row is a segment; the columns are the shared metric set measured within it. Opportunity is a tunable, goal-dependent prioritization score (tuned to enrollment here) — the single most decision-relevant column, computed in the app.
Size vs. clinical need
↑ more patients · → more clinical need
How to read it: up = a bigger population; right = sicker / less controlled. Each labelled corner tells you what a bubble’s position means. The right corner isn’t automatically “the answer” — a large, low-need segment (top-left) can be the best enrollment play, while a small, high-need segment (bottom-right) is the best clinical play.
Reading guide

Clinical Triggers — the 6 RPM-qualifying flags (Hypertension)

Addressable clinical triggers
where clinical need meets reachability
The engine behind the acuity routing above: count how many of these 6 flags a patient carries → 0 = ACCESS, 1 = Edge, ≥2 = RPM. Overlap-first, so a patient can carry several.

Glossary

Segment
A slice of the Market grouped by some shared trait. You study it to decide where to focus — you don't act on it directly (that's a Target list, built later).
Segment axis
The trait you slice by — the rows of the table (care model, complexity, ACO overlap…). Switching the axis keeps the same columns; only the rows change.
Acuity
How clinically severe or unstable a patient is right now. High acuity = needs intensive care. Not the same as whether they already have a care manager.
ACCESS / Edge / RPM Candidate
The three acuity buckets, set by how many of the 6 clinical flags a patient has: 0 = ACCESS (stable, light-touch), 1 = Edge (could go either way), ≥2 = RPM (acute, intensive).
Opportunity
A 0–100 prioritization score computed in the app — not a data-pipeline number. Its meaning depends on your goal; here it's tuned for enrollment (big, easy-to-engage = high).
PDC (adherence)
Proportion of Days Covered — the share of days a patient actually had their medication on hand. ≥0.8 (80%) = "adherent" (the CMS Star standard). From Part D pharmacy claims.
BP control
Whether blood pressure is on-target (<140/90). Pulled from HEDIS (Medicare Advantage) and CPT-II codes (fee-for-service). "Unknown" where neither source reports it.
VBC overlap
Share of a segment attributed to an ACO (a value-based-care org — SSP or REACH). High overlap = already "spoken for," harder to win.
Penetration
Your patients ÷ the market patients, within a segment (Panel ÷ Market). Grow-only — needs your uploaded patient list. Shows where you're strong vs. absent.
Suppression
A CMS privacy rule: any count under 11 patients is hidden (shown as "—") so no small group can be re-identified. It's why some cells are blank.
Cube
One pre-aggregated dataset keyed by several traits at once. Any segmentation view is a roll-up of it — so one dataset can serve many views.
Panel
Your own enrolled/attributed patients — the list you upload in Grow mode, matched against the market.