Skip to main content
. Author manuscript; available in PMC: 2016 May 25.
Published in final edited form as: Med Care. 2010 Apr;48(4):372–379. doi: 10.1097/MLR.0b013e3181ca404e

Table 3.

The impact of the publication of scientific evidence and the PA policy on the immediate probability of gabapentin use and the time trend of gabapentin use

Immediate Impact Time Trend
Predicted probability* Slope of logistic GEE model Change in predicted probability after 1 year**
Without event With event Change Without event With event Without event With event Change
Evidence 12.5% 13.6% 1.1% 0.14 0 4.6% −0.4% −5.0%
PA policy 12.4% 7.0% −5.4% 0 −0.08 −0.2% −2.3% −2.1%
*

Calculated as the difference between the predicted probability if the event were not to happen and the impact of the regression with the coefficients for the time trend turned off.

**

Calculated as the difference between the predicted probability after one year with the coefficients for the immediate impact turned off