Abstract

background

Participant self-report is the primary method of collecting data during actual use studies (AUS). Medication use diaries suffer from missing data and reactivity (the potential to influence the behavior being measured). Time sampling minimizes reactivity and records study product use during sample periods. We compared the results of two AUSs at the daily and weekly dosing levels for a once-daily oral medicine in men, using both methods.

methods

Time sampling data came from periodic telephone interviews about the use of study product during the previous 7 days. In AUS1, men were randomly assigned to receive an electronic diary to record daily use in addition to time sampling (Group 1) or undergo only time sampling (Group 2). In AUS2, men completed paper diaries in addition to time sampling.

results

Percentages of agreement between both methods were calculated for Days 1–7 (ranges: AUS1=60.7%–69.8%; AUS2=44.8%–75.5%). For values that did not match, percentages were quite low (ranges: AUS1=1.4%–5.8%; AUS2=0.8%–3.5%). Neither of the studies reported a downward trend in the interview periods. Percentages of agreements were affected by the number of missing values from the diary data (AUS1: Diary=24.4%–35.2% vs. time sampling <1.0%; AUS2: Diary=16.2%–27.1% vs. time sampling <1.0%). Analysis using non-missing data demonstrated a higher percentage of agreement between time sampling and diary data, suggesting similar results for both methods (day range: AUS1 91.8%–98.0%; AUS2 94.7%–98.4%).

conclusions

For products with a simple dosing regimen, time sampling is simple, has far fewer missing data points than diary data, is less intrusive, and thus may be better than diaries for an AUS.