How to improve PDC for medication adherence

To improve PDC scores, work your adherence list in fail-date order, reach patients weeks before they run out, and hold a two-way conversation that finds the actual barrier: cost, side effects, confusion or a ride. Route clinical questions to a pharmacist, enroll chronic patients in med sync, and track every outcome so nobody slips through.

What PDC measures, exactly

PDC stands for proportion of days covered. The Pharmacy Quality Alliance (PQA) defines it as the percentage of days in the measurement period covered by prescription claims for the same drug or drug class. A patient counts as adherent at 80% or higher. For antiretrovirals the bar is 90%. See the PDC glossary entry for a shorter definition.

The spec has four rules worth memorizing:

• Two fills. The patient needs at least two fills on different dates to enter the measure.
• The treatment period starts at the first fill. It runs from that first fill date to the end of the year, or to death or a gap in enrollment.
• 91 days. The first fill must land at least 91 days before the period ends. A patient who starts in mid-October is not in the measure that year.
• Exclusions. In the PQA diabetes specification, patients in hospice, patients with end-stage renal disease (ESRD), and anyone with an insulin prescription are excluded.

That second rule matters most. A patient who first fills in March has a shorter period, and a smaller budget of missed days, than one who started in January. The fail date guide walks through the math with a worked example.

Why 2027 work counts more than 2026 work

Medicare Part D has three adherence measures: diabetes medications, RAS antagonists for blood pressure, and statins. CMS lists all three at weight 3 in the 2027 Star Ratings. Weight 3 means each one counts three times as much as a standard process measure.

Two changes follow. First, CMS finalized sociodemographic status (SDS) risk adjustment for these measures starting with the 2028 Star Ratings. SDS adjustment accounts for factors like low-income subsidy status when scoring plans. Second, because that counts as a substantive change, consultant analyses from Wakely and RISE put the adherence measures at weight 1 for 2028 Stars before they return to weight 3 for 2029.

Star Ratings lag the data by about two years. The fills you chase in 2027 feed the 2029 ratings, where the triple weight comes back. For plans and the pharmacies in their networks, 2027 is the year to build a process that holds up. For more on how ratings drive plan revenue, read our Star Ratings overview.

Non-adherence is usually a barrier, not a memory problem

Most adherence programs are built like the patient forgot. Send a text, leave a voicemail, repeat. That works for the patient who did forget. It does nothing for the others.

Listen to the patients who miss refills and the reasons sound like this:

Each of those has a fix. Cost may point to the Medicare Prescription Payment Plan, a generic, or a 90-day fill. Side effects need a pharmacist and often the prescriber. Logistics may need delivery or med sync. None of them gets fixed by a reminder that only announces the gap.

The playbook

Here is the order of operations we'd run for a community pharmacy or a plan-side adherence team.

Two notes on the list itself. First, clean it. Patients who already filled at another pharmacy, switched drugs, or were discontinued by the prescriber still show up on claims-based lists. Mark them and move on. Second, track outcomes by reason, not just by "reached" or "not reached." A month of reasons shows you whether you have a cost problem or a side-effect problem.

What this looks like with an AI workspace

Rivvi is a HIPAA-compliant AI workspace. Staff upload files, such as a plan's adherence report or a fill-history export, and ask it to sort, flag and summarize. On the Team plan the same workspace places calls and texts, so the list and the outreach live together.

For adherence, that means the AI calls patients in fail-date order and asks how the medication is going. It records the barrier. When a patient has a clinical question, it warm-transfers them live to a pharmacist or leaves a task with the details. It speaks Spanish and 30 more languages. Opt-outs are honored on every path. Outcomes are recorded in Rivvi and pushed to a connected CRM such as HubSpot or Salesforce.

Southeast Medical Group ran this model and reported results in August 2026:

The pharmacist number is the one to sit with. One pharmacist reaching about 200 patients by hand is a ceiling, not a staffing problem you can hire out of. With the AI doing the first conversation, one pharmacist reached 30,000 patients.

See how this fits a pharmacy workflow on the medication adherence page, or compare options in our adherence software roundup.

A quick self-check

Before your next adherence cycle, answer these:

1. Do you know each patient's fail date, or only that they're "late"?
2. Does outreach start four or more weeks before the fail date?
3. Can a patient answer your outreach and get a pharmacist, or only hear a message?
4. Do you record why a patient missed, in categories you can count?
5. Are your chronic, multi-drug patients on med sync?

If you answered no to two or more, start with the first one. Sorting by fail date costs nothing and changes who gets called today. For pharmacy-specific workflows, see Rivvi for pharmacies.

Sources

• PQA, Measures resources (PDC definition and thresholds)
• PQA PDC-DR specification (via Florida Blue)
• Johns Hopkins Health Plans, Medication Adherence for Diabetes
• CMS, 2027 Star Ratings measures and weights
• Federal Register, CY2024 MA and Part D final rule (SDS risk adjustment), Apr 12, 2023
• Wakely, 2027 Final Rule Star Ratings changes
• RISE, Star Ratings measure set shrinks
• Medicare.gov, Costs for Medicare drug coverage