Non-Adherence to Prescribed Therapy is a Major Obstacle for Population PK/PD Studies

PDF of poster

Background: The extent of non-adherence to prescribed therapy in ambulatory patients is too often overlooked and ignored in the analysis of clinical trials. From clinical studies whose sponsors have concurred with entering their anonymized data into a common archive, we have recently finished the assembly and begun the analysis of data on 15214 ambulatory patients whose dosing histories during studies of varying lengths have been electronically compiled.

Methods: Electronic Medication Event Monitors were used to record the times and dates of package entry during the course of 87 drug studies performed between 1990 and 2004.  Chapter headings in the British National Formulary served to categorize fields of treatment. 

Results: Study durations ranged from 30 to 1400 days. Patterns of deviation from prescribed dosing regimens varied widely, but were almost entirely markedly skewed toward longer dosing intervals than prescribed, i.e. under-dosing, in every field of treatment.

On average, only during 60% of the study days was the prescribed dosing regimen executed correctly. In studies continuing beyond 6 [12] months, almost 30% [40%] of trial participants had stopped taking the trial medication, despite a prescription that called for continued taking of the drug. Drug holidays (3 or more consecutive days without drug intake) occur at least once a year in 50% of patients.  Holidays occur monthly in 2% of the patients, and quarterly in an additional 10% of patients. With QD BID regimens, 20% [40%] of doses were not taken on schedule. 

Conclusions: Underdosing, drug holidays, and early cessation of dosing are common features of clinical trials, and likely are frequent sources of low response and high variability in response to the protocol-specified dosing regimen. These dosing errors are usually grossly underestimated by counting returned, untaken dosage forms or by asking patients to fill out diaries or questionnaires. In the presence of non-adherence, missing or biased information on the dosage schedule leads to biased PK-PD analysis [1,2] and more unfortunately it forces one to discard up to 35 % of collected PK data [3]. The above findings highlight the needs for reliable assessment of dosing histories in clinical trials.

References :
[1]  Girard P, Sheiner LB, Kastrissios H, Blaschke TF (1996). Do we need full compliance data for population pharmacokinetic analysis ?. Journal of Pharmacokinetics and Biopharmaceutics, 24(3), 265-282.
[2] Vrijens B, Goetghebeur E (2004).  Electronic monitoring of variation in drug intakes can reduce bias and improve precision on pharmacokinetic/pharmacodynamic population studies. Statistics in Medicine, 23:531-544.
[3]  Lu J, Gries JM, Verotta D, Sheiner LB (2001). Selecting reliable pharmacokinetic data for explanatory analyses of clinical trials in the presence of possible noncompliance. J Pharmacokinet Pharmacodyn. 2001 Aug;28(4):343-62.