Change to Opioid study EMA sampling


We received the following request from Dr. Finan, with approval from Dr. Kumar and Sameer:

Our team here has been discussing the opioid reports we’ve been getting from the EMA and whether we have the optimal sampling technique down.

My primary concern with our current approach is that assessing opioid use both randomly and through event contingent entries can create confusion and may not yield the data we want. For example, some participants are only reporting about opioid use through the random prompts and not using the pain medication button. Another participant did the opposite. I think it would be more straight forward to go with just the Pain Medication button and double-down on training the participant to use it.

I suggest the following:
1) Remove opioid use questions from the random prompts.
2) Retain the event contingent “Pain Medication” button as is, and we will focus on that critically in our participant training
3) Add an opioid use item at the fixed end-of-day prompt, in which we ask the participant to tell us how many pills the took over the course of the entire day. (this would give us a check on their use of the Pain Medication button, so Shriya can call them to reinforce the training if she sees a big discrepancy)

Can you weigh in on whether you agree with this type of change? If so, would it be possible to make these changes prior to our next participant, who is coming in Wednesday, May 8th?


Monowar Hossain


Shahin Samiei