A New Statistical Model for Improving Contraceptive-device Trials

A USAID-supported study from Family Health International highlights the complications inherent in evaluating the effectiveness of new contraceptive devices — and proposes a new statistical model to help overcome these challenges.

The study evaluated data from a 2003 trial of the eZon plastic condom. Although the data suggest that the ineffectiveness of a new device can reflect an actual failure on the part of the device, it can also result from other factors such as:

• Learning effects: Before a participant becomes skilled in the proper use of a new device, he or she may experience a high failure rate.
• Noncompliance: A participant may not be comfortable with the technique needed to use the device, so he or she fails to follow the study protocol.
• Nonrandom dropout: Participants in a new-device arm of a study leave the study at higher rates than participants in the control arm, possibly because they feel uncomfortable or feel incompetent with the new device.

The author used the data from the condom trial to propose the new statistical model, which assesses how learning effects, noncompliance, and nonrandom dropout influence the effectiveness of new devices during clinical trials. The proposed model could help researchers better understand the results of current contraceptive-device trials and improve the design of future trials.

To read more about this topic, click here.

Source
Chen, Pai-Lien. Assessing learning effects and nonrandom dropout in a contraceptive device trial. J Biopharm Stat 2008;18(2):382-392.