Signature
MPR_capped = min(MPR, 1)
| Inputs | Definition | Unit |
|---|---|---|
MPR | Uncapped ratio from any variant, HE-FM-MPR-001 to HE-FM-MPR-003 | proportion, zero or above |
MPR_capped | Medication possession ratio truncated at 1 | proportion, at most 1 |
|---|
Function
Medication possession ratio from summed days' supply function
Maps a patient's dispensing records for one medicine, each a fill date and a days' supply, to the medication possession ratio (MPR): the days' supply dispensed divided by a count of days. The variants differ in which fills enter the numerator and which days form the denominator: a fixed interval from the first dispensing to the end of the observation period, or a refill interval between the first and last dispensing, with or without the last fill. Because supply is summed without regard to overlap, an uncapped MPR can exceed 1, and it is often capped at 1. The records follow the notation of the Medication Possession Ratio article. The proportion of days covered, which counts unique covered days instead, is HE-FM-PDC-001.
Try this function
Implementations
Excel
Capped MPR from a named uncapped ratio cell
Excel takes the smaller of the named cell holding the uncapped ratio and 1.
=MIN(MPRUncapped,1)
Assumptions
Cap applied to each patient's ratio before averaging
The cap is applied to each patient's ratio, and the variant it is applied to is stated, because capping does not reconcile different denominators.
Worked examples
Capped fixed-interval MPR of the five-fill history
The fixed-interval MPR of about 1.167 for the article's fills is capped at 1.000, which reads as perfect adherence although 30 of the 180 days had no medicine.
MPR = 1.1667; MPR_capped = 1
Cap leaves an MPR below 1 unchanged
A patient who stops after the third fill has a fixed-interval MPR of 0.5, which the cap leaves unchanged (computed here for illustration).
MPR = 0.5; MPR_capped = 0.5
Common errors
Averaging uncapped MPRs
Among adults with asthma using inhaled corticosteroids, Vollmer and colleagues found means of 0.61 and 0.63 for two uncapped measures often called the MPR, with maxima of 50 and 100, against 0.55 and 0.52 once each was capped at 1. A few extreme values pull the uncapped mean upwards.
Capping the mean MPR instead of each patient's MPR
Capping a group mean of 0.61 leaves it at 0.61, while capping each patient's ratio first gave a mean of 0.55 in the same data, so the skew from oversupply stays in the result (computed here for illustration).
Sources
Capped measures defined as the minimum of the uncapped ratio and 1
Vollmer WM, Xu M, Feldstein A, Smith D, Waterbury A, Rand C. Comparison of pharmacy-based measures of medication adherence. BMC Health Services Research. 2012;12:155. Methods and Table 2: CMA3 = minimum (CMA1, 1) and CMA4 = minimum (CMA2, 1); CMA1 and CMA2 are often called the MPR, with means of 0.61 and 0.63, 95th percentiles of 1.24 and 1.33 and maxima of 50 and 100, against means of 0.55 and 0.52 when capped.
Truncation of the MPR at 1.0 after early refilling
Sikka R, Xia F, Aubert RE. Estimating medication persistency using administrative claims data. American Journal of Managed Care. 2005;11(7):449-457. Introduction: early refilling leads to an MPR of more than 1.0, which is often truncated at the maximum value of 1.0.
Canonical Identity
Stable URI · Machine-readable · Resolvable · CC BY 4.0