Drug Payer & Channel Explorer

2023

Search and select one or more drugs from the 532 unique medicines in the MEPS 2023 prescribed medicines file after excluding missing drug names (187,060 fills, 11,836 patients). Selected drugs are combined into one view — who paid, and where the prescription was filled.

Coverage, in AHRQ's own words: "Each record represents one household-reported fill or refill of a prescribed medicine that was obtained during calendar year 2023 at any retail pharmacy, including mail-order or on-line." And on what's out of scope: "The HC does not collect information on drugs administered in healthcare settings" — respondents report drugs prescribed during a doctor's-office, clinic, ER, or inpatient visit only if the drug was subsequently filled, not drugs given during the visit itself. That's why IV-infused oncology biologics (Herceptin/trastuzumab, Keytruda/pembrolizumab, Perjeta/pertuzumab, Avastin/bevacizumab, Opdivo/nivolumab — all 0 records here) and other in-office-administered products (Nexplanon, IUDs) don't appear: they're billed as part of the clinical visit, captured in MEPS's Office-Based or Outpatient event files instead. Sources: AHRQ MEPS HC-248A 2023 documentation §2.0; Methodology Report #37, "Household Component" §.

Therapeutic class

Sub-class

Names are generic / molecule names, not brand names — e.g. search atorvastatin for Lipitor, escitalopram for Lexapro, levothyroxine for Synthroid. Per AHRQ: "This drug name [RXDRGNAM] is the generic name of the drug most commonly used by prescribing physicians. It is supplied by the Multum Lexicon database." A generic name can cover multiple brand-name and store-brand products. A handful of entries in this list (e.g. "Hormones/Hormone Modifiers") are not real drug names at all — AHRQ masks orphan drugs and drugs estimated to be used by fewer than 400,000 people nationally, for confidentiality, and recodes them down to their therapeutic class instead.

View by

Dollar amountTotal dollars in that bucket. For payer, this is already exclusive by construction — every dollar belongs to exactly one payer.

FillsCount of prescription fills. For channel and dosage form, a fill has exactly one value, so this is a plain count. For payer, a fill can be split across multiple payers (see the red note under "Payer split" below).

Days suppliedTotal days of medication supplied in that bucket, summed from RXDAYSUP across fills (excluding missing/non-numeric values).

Payer split — proportional attribution

⚠ Payer fills/days are attributed, not exact. A single fill can be paid by more than one payer, so — unlike channel and dosage form, where every fill has exactly one value — payer fills/days use proportional attribution: split by each payer's dollar share. A fill where Medicare pays 80% and the patient pays 20% counts as 0.8 fills for Medicare and 0.2 for the patient, so the three metrics each sum to exactly 100% with no double-counting. The "Fills" / "Days supplied" toggle above is shared across all three charts, but only here does it mean attributed fills/days.

MedicareFederal health insurance for adults 65 and older. Also covers younger people with long-term disabilities (after 24 months on SSDI), end-stage renal disease (ESRD), or ALS.

MedicaidJoint federal–state program covering low-income individuals and families. Eligibility and benefits vary by state.

PrivatePrimarily employer-sponsored group plans, plus individually purchased marketplace plans.

Patient OOPOut-of-pocket payments made directly by the patient or family — copays, deductibles, and costs for uncovered services.

VA/CHAMPVADepartment of Veterans Affairs coverage for eligible veterans. CHAMPVA extends to spouses and dependents of veterans with service-related disabilities.

TricareHealth coverage for active-duty military, National Guard and Reserve members, retirees, and their families.

Other federalOther federal government programs, e.g. Indian Health Service or federal correctional facilities.

State/localState and local government health programs outside of Medicaid.

Workers compWorkers' compensation insurance covering treatment for work-related injury or illness.

OtherMiscellaneous payer sources not captured by the categories above.

PAPFills with $0 total expenditure — patient assistance program or fully waived cost, not attributed to any payer.

Pharmacy channel — PHARTP1

Drug storeRetail pharmacy chain or independent drugstore — the default channel for most routine oral medications.

In another storeNon-drug-store retail location with a pharmacy counter, e.g. a grocery store or big-box retailer.

Mail-orderPrescription shipped directly to the patient, typically used for long-term maintenance medications.

In HMO/clinic/hospitalDispensed on-site at a clinic, hospital, or HMO facility — common for injectables, infusions, and provider-administered drugs.

OnlinePrescription ordered and fulfilled through an online pharmacy.

PHARTP1 = type of pharmacy provider, first-listed source per fill. AHRQ's own caveat: when a household reports using more than one type of pharmacy over the year, "there is no link in the survey or in the data file enabling analysts to know the type of pharmacy from which a specific prescription was obtained if multiple pharmacies are listed" — so channel attribution is somewhat noisier for patients who use more than one pharmacy type. Source: AHRQ MEPS HC-248A 2023 documentation, "Type of Pharmacy" §.

Dosage form — RXFORM

Official AHRQ decode table. Labels shown here (e.g. "Tablets" for TABS) come verbatim from Appendix 1, "Definitions for RXFORM, Dosage Form," of AHRQ's own HC-248A documentation (558 codes). Some codes have no label because AHRQ's own table marks them "no definition for the dosage form" — for example SOPN, SOAJ, and PSKT are undefined even in the official source, not a gap in our mapping. The raw code is always shown alongside the label so you can verify it against the source yourself. Source: AHRQ MEPS HC-248A 2023 documentation, Appendix 1.

Multi-drug patient count. When more than one drug is selected, the unique-patient count is deduplicated (a patient on two selected drugs is counted once) — it is not the sum of each drug's patient count.

Source. AHRQ, 2023 Medical Expenditure Panel Survey, Prescribed Medicines File (HC-248A) — official documentation, codebook, and Methodology Report #37 on data collection. 192,275 total records (matches AHRQ's published count exactly); 190,730 have a positive person-level weight (PERWT23F). All statistics on this page are unweighted counts of the 187,060 fills with an imputed drug name (5,215 fills with unimputed/missing RXDRGNAM excluded from the drug list). Days-supplied is excluded from day-based metrics for fills with missing (-8/-7) or non-numeric (999 = "taken as needed") RXDAYSUP — per AHRQ, no consistency edits were applied between quantity and days supplied, and no edits were made for very high values. Since the 2017 data, MEPS stopped asking households to report drug payments directly — all expenditure and payer-source figures on this page come from the Pharmacy Component (pharmacy records), not household recall.