When opioid therapy is initiated for a new pain condition, it may be unknown whether the pain will persist beyond the time of tissue healing. The aim of this study was to determine the prevalence of prescription patterns indicating persistent and/or problematic opioid use in a cohort of opioid-naive patients starting therapy with weak opioids. Data were drawn from the nationwide Norwegian Prescription Database. The study population was all new users of opioids receiving prescriptions of a weak opioid in 2005 for nonmalignant pain. This cohort was followed until December 2008. In order to be classified as having probable problematic opioid use, patients had to meet all of the following criteria: received opioids at least once every year from 2005 to 2008 and in 2008; (1) were dispensed more than 365 defined daily doses (DDDs) of opioids; (2) received opioid prescriptions from more than 3 doctors; and (3) were dispensed more than 100 DDDs of benzodiazepines. There were 245,006 persons who were new users of weak opioids in 2005 (216,902 codeine, 26,326 tramadol, 1778 dextropropoxyphene). There were 17,252 (7% of new users) who received a prescription for opioids at least once each of the 3 following years. Of these subjects, 686 patients were dispensed more than 365 DDDs of opioids in 2008 and are probably persistent users. There were 191 subjects who met our criteria for probable problematic opioid use. In a cohort of new opioid users who started treatment with weak opioids, only 0.3% and 0.08% developed prescription patterns indicating persistent opioid use and problematic opioid use, respectively.