Threshold

Recommendation:

• State whether the PDC was analysed as a dichotomous or continuous variable.

• If the PDC was dichotomized, provide a rationale for selecting this threshold.

• Consider conducting a sensitivity analysis with PDC analysed as a continuous variable, or dichotomized at an alternative cut-off informed from the data, literature or other method.

Explanation:

While a threshold of 80% is commonly used in the literature to define high adherence, there is increasing evidence to indicate this threshold is arbitrary and may not be the most appropriate cut-off for classifying patients as "adherent". Where dichotomisation of the PDC is absolutely necessary, it is recommended that researchers use alternative thresholds at higher (or lower) thresholds to demonstrate robustness of findings. 

Eligibility criteria for inclusion in sample

Recommendation:

• Define the eligibility criteria for assessing the PDC, e.g. minimum number of scripts required to be filled within a period.

Explanation:

For example, in many medication adherence studies, researchers limit the analytical sample to only include patients who filled two or more prescriptions within a certain timeframe. Limiting analyses to patients with ≥2 dispensings may help to focus on patients who initiated therapy for long-term use (e.g., excluding any trial prescriptions or short/acute courses of treatment).

Numerator and denominator

Recommendation:

• Define the numerator and denominator for the PDC, e.g. The PDC was defined as the total number of days with at least 1 medication available (numerator) in the 1-year period following hospital discharge (denominator).

Explanation:

The numerator/denominator typically commences from the first medication supply date in the observation period.

In studies examining medication adherence following an acute hospitalisation, it may be appropriate to commence the denominator follow-up period from the date of hospital discharge following hospitalisation. This helps to understand medication adherence in the early post-discharge transitional period. 

Survival

Recommendation:

• State whether the PDC was assessed among individuals surviving the entire measurement period or whether the PDC was measured until the date of death for those who died during the observation period.

Explanation:

To minimise selection bias, it is often preferred to include all participants and measure medication adherence until the date of death. The follow-up period can be truncated to focus on the number of days a person was alive.

Pre-supply

Recommendation:

• State whether analyses were limited to new users or how presupply was handled if previous users of medication were included.

Explanation:

When measuring medication adherence from a fixed (e.g., 1 Jan 2025) or event-related date (i.e., date of hospital discharge), and prevalent users of medication are included, it may be necessary to account for any medication pre-supply available at the start of the follow-up period. 

In-hospital supply

Recommendation:

• State whether information on in-hospital medication dispensing was available.

• If unavailable, describe how periods of time spent in hospital were handled, e.g. periods where an individual was admitted to hospital were excluded from both the PDC numerator and denominator.

Explanation:

In many jurisdictions, data on medications supplied in hospital are not captured in the medication dispensing datasets. Consequently, patients who are re-hospitalised during the follow-up period may be misclassified as non-adherent (despite receiving a separate supply of medication while in hospital). To overcome this limitation with data availability, it may be necessary to either: 

1) exclude hospitalised days from the numerator and denominator; or

2) add hospitalised days to the numerator (i.e., assumes the patient received separate medication supplies while in hospital, and community-based supplies will be re-commenced on hospital discharge)

Dosing information

Recommendation:

• State whether dosing information was available in the data or imputed.

• If applicable, describe the approach used to impute doses and the validity and/or limitations of this method.

Explanation:

In many jurisdictions, information on the prescribed daily dose (i.e., quantity of medications to be taken per day) is not collected or available within the medication dispensing dataset. This information is required to reliably determine periods of medication exposure. Where surrogate daily doses cannot be reliably imputed (e.g., from prescribing guidelines), a data-driven approach to determining daily doses may be required. This commonly involves imputing the 75th (or 80th) percentile of time taken for patients to return for a medication refill.

For more information, see: Ung D, Dalli LL, Lopez D, et al. Assuming one dose per day yields a similar estimate of medication adherence in patients with stroke: An exploratory analysis using linked registry data. Br J Clin Pharmacol. 2021;87(3):1089-1097. doi.org/10.1111/bcp.14468

Early refills

Recommendation:

• Describe how overlapping supplies due to early refills of the same medication (i.e. stockpiling) were handled, e.g. carry-over was granted for early refills of the same drug.

Explanation:

A benefit of the PDC method is the ability to account for early refills of the same medication. This is particularly important in jurisdictions where stockpiling of medications may occur (e.g., in Australia, prescriptions are valid for 12 months and typically include 1 original and 5 repeats. Some patients may seek approval to have their entire prescription dispensed at once to minimise visits to the pharmacy). It is important to account for overlapping supplies of the same medication when calculating the PDC.

Switching

Recommendation:

• Describe how overlapping supplies of different medications within the same therapeutic class were handled (i.e. therapeutic switches), e.g. carry-over was not granted for therapeutic switches (e.g. switching from simvastatin to atorvastatin).

Explanation:

When accounting for overlapping supplies of medications, it is important to consider whether these medications were the same or different. Overlapping supplies of two different medicines within the same therapeutic class (i.e., simvastatin and atorvastatin) may indicate that a therapeutic switch has occurred. It is unlikely that patients would be taking both medications concurrently, and it may be reasonable to assume the first medication has been discontinued on receipt of the new medication in the same class.