This week the FDA named Dexcom the first participant in its TEMPO pilot — the first serious attempt to make device oversight and Medicare payment move on one clock instead of two. Here's why that's a big deal: twelve years of Medicare claims for continuous glucose monitoring show that adoption never moved when the sensors got better. It moved when the coverage rules changed. Press replay and watch the policy dates bend the curve.
CPT 95251 is the claim a clinician files for interpreting a patient's CGM data — the professional footprint of CGM in traditional Medicare. In 2013 it was a niche endocrinology service. By 2024 it was 14x bigger, while its clinic-owned sibling (95250, the "loaner sensor" service) peaked in 2018 and fell back below its 2013 level. The patient-owned sensor won — and the handoff date was set by CMS, not by the technology.
368,447
patients with CGM interpretation billed, 2024 (95251)
14x
growth in interpretation patients since 2013
15,426
clinicians billing 95251 in 2024 (vs 3,226 in 2013)
−66%
clinic-owned CGM (95250) patients since its 2018 peak
The animated record
Every kink is a coverage decision
Two lines: patients whose personal CGM data a clinician interpreted (95251) and patients set up on a clinic-owned loaner CGM (95250). The flags are payment-policy events. Watch what happens to each line as the flags go by — then ask yourself which mattered more, the sensor or the rule.
Medicare FFS patients with CGM professional claims, by year · national
95251 · personal CGM, interpretation95250 · clinic-owned CGM, setup/loaner
2013
year
26,397
personal-CGM interpretation patients
25,061
clinic-owned CGM patients
Hover any year for exact counts. Counts are distinct FFS beneficiaries per code per year; CMS suppresses cells under 11.
Jan 2017CMS Ruling 1682-R: therapeutic CGM classified as durable medical equipment — Medicare coverage of personal CGM begins. The curve's first real bend.
2018CPT overhaul: 95249 created for patient-owned CGM setup; interpretation of personal devices now cleanly billable. Clinic-owned CGM peaks this year — then starts dying.
Jul 2021Requirement dropped: Medicare removes the 4-finger-sticks-per-day testing rule for CGM eligibility. Growth steepens.
Apr 2023LCD expansion: coverage extends to anyone on insulin — any regimen — and to patients with problematic hypoglycemia. The steepest segment on the chart follows.
Jul 2026TEMPO begins: FDA names Dexcom participant #1; CMS's ACCESS model supplies the payment path from day one. For the first time, the two clocks are meant to tick together.
What this chart is not telling you
Read it with one eyebrow up
This is the professional-fee sliver, not CGM. These codes paid clinicians about $25M in 2024 — the sensors themselves bill through the DME channel (K0554, A4239), a far larger dollar flow that isn't on this chart. And it's traditional Medicare only: the majority of beneficiaries now sit in Medicare Advantage, whose CGM use is invisible here. So the level of each line understates reality — but the timing of the bends is exactly the point, and the bends land on rule changes, not product launches.
The critical read: Dexcom's G4 arrived in 2012, G5 in 2015, G6 in 2018 — steady sensor progress across the whole window. If technology drove adoption, this curve would be smooth. It isn't. It's a staircase, and every stair has a Federal Register citation. A clinical AI tool with FDA clearance and no payment path inherits the flat part of this staircase — which is why a pilot that starts with the payment model attached is worth your attention.
A decade of CGM says the binding constraint on clinical technology isn't clearance and it isn't capability — it's the coverage rule. TEMPO is the first pilot designed as if someone at FDA had stared at this exact chart. 🔮 If it works, "TEMPO-eligible" shows up on pitch decks within 18 months, the way "FDA-cleared" does now.