The chart appears to show that OxyContin doses peak quickly and then plateau in the blood stream, providing constant levels of pain relief. It was part of OxyContin's labelling when Michael Friedman and Howard Udell were Purdue's CEO and chief counsel, respectively.
But sharp-eyed observers will note that the chart is in log scale, not linear scale. In a regular linear scale graph -- the type of scale everyone is used to seeing on a daily basis in everything from the Dow Jones Index to your quarterly sales numbers -- each point on the vertical axis represents the same interval from the previous point; i.e. 1, 2, 3, 4, etc.
In log scale, however, each point on the vertical axis is a multiple of the previous point, i.e. 1, 10, 100, 1000, etc. A log scale can thus mask changes in a curve because large differences between points are concealed within seemingly "regular" intervals. If the curve in the log graph is shown in regular linear scale, it looks like this:
Clearly, the OxyContin levels peak quickly, but then they decline quickly as well. Which is one of the reasons that anyone taking OxyContin often wants to take more -- within a few hours of a new dose, it's already wearing off.
According to Philip Thomas, an attorney petitioning the FDA to change OxyContin's labelling, the log chart was used by Purdue's sales reps to persuade doctors that OxyContin would not be that addictive because there was no sharp "comedown" for its users. Reps and docs didn't fully understand that the curve was shown on a log scale, not a regular scale. (Download the petition here.) Purdue's current management opposes Phillips' petition. The plea agreement signed by Friedman and Udell does not make it clear which executives at Purdue signed off on the chart. (Download the "agreed statement of facts" in the Purdue case here.) The Department of Health and Human Services later banned the pair from doing business with the government based on their conviction for promotion of misbranded drugs. Friedman and Udell have sued to overturn the ban.
Wouldn't it be interesting to know who, exactly, had the final responsibility in signing off on that chart?