Proportion of Days Covered

Proportion of Days Covered, the 80% Threshold and the Contrast with MPR

The proportion of days covered (PDC) sits behind the Medicare Part D adherence ratings in the United States and is widely used in database research, so a health economist meets it as a model input, as the outcome of adherence interventions and as a quality measure that carries payment. A PDC of 0.85 means that pharmacy records show medicine on hand on 85% of the days in the period measured. The calculation has to handle overlapping refills and hospital stays. The usual 80% cut-off is a convention, and PDC differs from the medication possession ratio (MPR).

A measure of medication availability built from dispensing data

PDC infers medication adherence from pharmacy dispensing or claims records. Each record gives a fill date and a days' supply, so the data show when medicine was collected and how long it should last, not when it was taken. A standard set of definitions for database studies describes PDC and MPR as the two most commonly used possession measures of secondary adherence, meaning whether a patient keeps receiving refills as prescribed once the first prescription has been dispensed.

The Pharmacy Quality Alliance (PQA), which develops measures of the quality of medication use, chose PDC as its standard method after an expert workgroup first convened in 2006 reviewed the literature and tested draft specifications. By 2013 PDC as operationalised by PQA had been endorsed by the National Quality Forum, and the Medicare Part D adherence measures in the Centers for Medicare & Medicaid Services (CMS) Star Ratings are adapted from it.

Calculating PDC step by step

The calculation counts unique calendar days, not days' supply. Each fill is laid out as a run of dates, and a day counts once however many fills cover it. In symbols, PDC is written as follows.

$$PDC = \frac{C}{T}$$

where $C$ is the number of days within the observation period on which at least one fill provides medicine, after any shifting of overlapping supply, and $T$ is the number of days in the observation period, counted from the index date to the end date inclusive. Because $C$ counts only days inside the period, PDC cannot exceed 1.

The PQA method, as written in its paper on PDC and in the CMS 2026 Star Ratings technical notes, runs as follows.

1. Set the index date. The index date is the first fill of the target medicine in the measurement period; CMS calls it the index prescription start date.

2. Set the end of the observation period. For PQA and CMS the period runs from the index date to the end of the measurement year, disenrolment or death, whichever comes first. A patient who stops refilling is still followed to that end date, so stopping treatment lowers the PDC. CMS also requires at least two fills on different dates and a treatment period of at least 91 days.

3. Map each fill to calendar days. A 30-day supply collected on 1 March covers 1 to 30 March.

4. Shift overlapping fills of the same drug. When a refill of the same drug arrives before the previous supply has run out, its start date moves to the day after the previous supply ends, on the assumption that the patient finishes the old supply first.

5. Do not shift overlaps between different drugs in the class. Where the measure covers a class, such as all oral diabetes medicines, overlapping fills of different drugs are treated as taken together, and a day counts as covered if at least one drug in the class is available.

6. Adjust for inpatient and skilled nursing facility stays. CMS removes stay days from both numerator and denominator and moves supply that overlaps a stay to uncovered days after discharge, assuming the facility supplied the medicine.

7. Count covered days and divide. Supply running past the end date is ignored.

Analyst choices that move the result

The steps above leave several decisions open, and each can change the figure. A 2022 scoping review of 76 articles measuring PDC for cardiovascular medicines found wide variation and proposed the TEN-SPIDERS reporting tool, which lists the parameters authors should state, from the threshold and eligibility criteria to early refills and switching.

Early refills were carried over in 45% of the articles reviewed but not reported in 37%. In-hospital supplies were not reported in 45%, and were otherwise handled in different ways, most often by excluding hospital days from both numerator and denominator. Where the prescribed dose was unavailable, 21% of all the articles assumed one unit a day.

How several related medicines are combined matters as much. In a cohort of 7,567 people taking two or more classes of oral diabetes medicine, concurrent adherence ranged from 35% to 95% depending on the definition. Counting a day as covered if any one class was available, close to the PQA class method, classed virtually all patients as adherent, whereas requiring a PDC of at least 80% for every class classed only 30% to 40% as adherent.

Worked example of five fills over 180 days

The figures are illustrative. A patient starts a once-daily medicine, and an analyst measures PDC over a 180-day period that begins at the first fill (day 1) and ends on day 180. The patient makes five fills of the same drug.

1. Shift the early refills. Fill 2 arrives on day 22 with 9 days of fill 1 left, so it starts on day 31. Fill 3 arrives on day 45 while shifted fill 2 lasts to day 60, so it starts on day 61. Fill 5 arrives on day 170 while fill 4 lasts to day 210, so it starts on day 211, after the period has ended.

2. Count covered days. Days 1 to 90 are covered (90 days), days 91 to 120 are not (30 days), and days 121 to 180 are covered (60 days).

3. Compute PDC.

$$PDC = \frac{150}{180} = 0.833$$

where 150 is the number of covered days and 180 the number of days in the period. The PDC is 83.3%, above an 80% threshold.

4. Without shifting. Merging overlaps instead, fills 1 to 3 cover days 1 to 74 and fill 4 covers days 121 to 180, giving 74 + 60 = 134 covered days, as shown below.

$$PDC_{\text{no shift}} = \frac{134}{180} = 0.744$$

where $PDC_{\text{no shift}}$ treats the stock from early refills as never used. At 74.4% the same patient falls below the threshold.

5. With a hospital stay. Suppose the patient was an inpatient on days 95 to 104, inside the gap. Following the CMS method, those 10 days leave the denominator; no supply overlapped the stay, so nothing is shifted and the numerator stays at 150 in the adjusted calculation below.

$$PDC_{\text{stay}} = \frac{150}{180 - 10} = \frac{150}{170} = 0.882$$

where $PDC_{\text{stay}}$ is the PDC after the stay adjustment, here 88.2%.

6. The MPR on the same fills. A medication possession ratio that sums all days' supply dispensed in the period gives

$$MPR = \frac{30 + 30 + 30 + 90 + 30}{180} = \frac{210}{180} = 1.167$$

where the numerator is total days' supply and 180 is the length of the period. The ratio of 116.7%, or 100% if capped, hides a 30-day gap that the PDC records.

7. What the numbers show. One fill history gives values from 74.4% to 116.7%, and the classification at 80% changes with the overlap rule alone, so any PDC used in a study, contract or model needs its rules stated.

Where the 80% threshold comes from

A PDC is often reported as a yes or no, adherent at or above a threshold, non-adherent below it. The standard definitions list PDC of at least 80% as a cut-point used by convention, only rarely with supporting evidence.

PQA states that it selected 0.8 on the basis of numerous studies of the link between adherence and health outcomes, as the level above which a patient can be considered highly adherent for most chronic medicine classes, with 90% chosen for antiretrovirals for HIV; its current measure list keeps both values. In the CMS 2026 Star Ratings, each of the three Part D adherence measures (diabetes medicines, renin angiotensin system antagonists and statins) reports the percentage of continuously enrolled members aged 18 and over with a PDC of 80% or higher.

In Arkansas Medicaid claims for five chronic conditions, the MPR and PDC cut-off that best predicted any-cause hospitalisation ranged from 0.63 to 0.89, which the authors described as an initial empirical basis for 0.80. The 2022 scoping review found that 61 of 73 articles used a threshold of at least 80%, and that only 56% of these gave a reason. It recommended stating the rationale for any threshold and running a sensitivity analysis with PDC as a continuous variable or with an alternative cut-off.

How PDC differs from the medication possession ratio

PDC and MPR both draw on fill dates and days' supply and are often confused. The main difference, as the standard definitions put it, is that PDC truncates any oversupply, whereas MPR allows values above 100%. PQA's analysis concluded that the two give nearly identical results for a single drug, and that PDC gives a more conservative estimate for drug classes with frequent switching and concurrent use.

Uses in economic evaluation, quality measurement and payment

Adherence as a model input. A 2026 microsimulation of 152,117 patients in Catalonia computed annual PDC from real-world data in the SIDIAP primary care records, defining optimal adherence as a PDC of at least 80%. Noting that most cardiovascular models assume full adherence, the authors used a state-transition model to compare observed adherence, full adherence and no treatment.

Cost-effectiveness of adherence interventions. In the same study, improving adherence cost between €8,946 and €12,614 per QALY gained. The threshold analysis turns a difference in adherence into a maximum price for an intervention. Moving patients from observed to full adherence would remain cost-effective at up to €5,411 to €5,914 per patient, at a willingness to pay of €25,000 per QALY. The analysis did not include the cost of the interventions themselves, which the authors listed as a limitation; a cost-effectiveness analysis of a real intervention would add that cost.

Quality measures and plan payment. The three Part D adherence measures carry a weighting value of 3 in the CMS 2026 Star Ratings, and contract Star Ratings are also used in determining quality bonus payment ratings.

Outcomes-based agreements. In an outcomes-based agreement between a manufacturer and a regional health plan comparing ticagrelor with clopidogrel after acute coronary syndrome, patients counted only if their PDC was at least 80% over a year or until a recurrent heart attack.

What a dispensing record cannot show

PDC is easy to calculate for whole insured populations, but it measures supply, not use, and its weaknesses follow from that.

• Collection is not consumption. A covered day means medicine was available, not that a dose was taken. A 1997 review judged refill measures useful in population studies when direct measurement of consumption is not feasible.
• The least adherent are missing. Patients who never fill a first prescription cannot have a PDC, and the standard definitions note that it systematically excludes the most non-adherent patients and is imprecise for those with one dispensing.
• Missing fills look like gaps. Prescriptions paid in cash or filled outside the database leave no claim, so the measure assumes the dispensing record is complete.
• Adherence and persistence are mixed. Because the period runs to a fixed end date, stopping treatment and missing doses both lower the PDC, and a very low value can reflect discontinuation. The ABC taxonomy separates initiation, implementation and discontinuation, and a single PDC does not distinguish the last two.
• Days' supply may be wrong. The numerator relies on the recorded days' supply or an imputed dose, and imputing one unit a day is less reliable for drugs whose dose varies.

Sources

• Centers for Medicare & Medicaid Services. Medicare 2026 Part C & D Star Ratings Technical Notes. Baltimore: CMS; updated 25 September 2025. Measures D08 to D10 and Attachment L.
• Choudhry NK, Shrank WH, Levin RL, Lee JL, Jan SA, Brookhart MA, et al. Measuring concurrent adherence to multiple related medications. American Journal of Managed Care. 2009;15(7):457-464.
• Dalli LL, Kilkenny MF, Arnet I, Sanfilippo FM, Cummings DM, Kapral MK, et al. Towards better reporting of the proportion of days covered method in cardiovascular medication adherence: a scoping review and new tool TEN-SPIDERS. British Journal of Clinical Pharmacology. 2022;88(10):4427-4442. doi:10.1111/bcp.15391
• Karve S, Cleves MA, Helm M, Hudson TJ, West DS, Martin BC. Good and poor adherence: optimal cut-point for adherence measures using administrative claims data. Current Medical Research and Opinion. 2009;25(9):2303-2310. doi:10.1185/03007990903126833
• Nau DP. Proportion of Days Covered (PDC) as a Preferred Method of Measuring Medication Adherence. Pharmacy Quality Alliance; undated, cited as 2011 by Raebel et al. (2013). Archived copy of 27 July 2018, accessed 2 October 2026.
• Peasah SK, Huang Y, Venditto J, Brekosky R, Belletti D, Campbell V, et al. Ticagrelor versus clopidogrel for recurrent myocardial infarction: an outcomes-based agreement. Exploratory Research in Clinical and Social Pharmacy. 2023;12:100347. doi:10.1016/j.rcsop.2023.100347
• Pharmacy Quality Alliance. Measures and Resources: medication adherence measures. Accessed 2 October 2026. https://www.pqa.org/measures/measures-resources/
• Raebel MA, Schmittdiel J, Karter AJ, Konieczny JL, Steiner JF. Standardizing terminology and definitions of medication adherence and persistence in research employing electronic databases. Medical Care. 2013;51(8 Suppl 3):S11-S21. doi:10.1097/MLR.0b013e31829b1d2a
• Sánchez-Viñas A, Wu R, Aznar-Lou I, Mihaylova B, Rubio-Valera M. Lifetime effects of adherence to cardiovascular and diabetes medications in Spain: a modelling study in a population cohort of 152,117 patients. PharmacoEconomics. 2026;44(7):851-863. doi:10.1007/s40273-026-01597-1
• Steiner JF, Prochazka AV. The assessment of refill compliance using pharmacy records: methods, validity, and applications. Journal of Clinical Epidemiology. 1997;50(1):105-116. doi:10.1016/s0895-4356(96)00268-5
• Vrijens B, De Geest S, Hughes DA, Przemyslaw K, Demonceau J, Ruppar T, et al. A new taxonomy for describing and defining adherence to medications. British Journal of Clinical Pharmacology. 2012;73(5):691-705. doi:10.1111/j.1365-2125.2012.04167.x