We hypothesize that increasing use of prescribed opioids for chronic pain in vulnerable patients with mental health disorders is the primary reason for escalating problems with opioid abuse and other negative health consequences of opioid use. This study will analyze 2002-2007 claims data from two complementary sources: 1] Highmark-Blue Shield private insurance plan (N= 3.5 million), 2] Arkansas Medicaid (N=660,000) We will study those individuals with the three most common chronic non-cancer pain conditions: back pain, joint pain, and headaches. In 2002, over 700,000 individuals from Highmark and over 100,000 individuals from Medicaid had a visit for one of these pain conditions. In 2002, 4.9% (n=34,819) of this Highmark chronic pain group and 12% (n=12,263) of this Medicaid chronic pain group were prescribed chronic opioids (>60 days/year), offering us over 250,000 person-years of observation of patients with chronic pain on chronic opioids . These datasets describe two groups who have geographical and demographic differences that are likely associated with different patterns of opioid use and different risks for opioid abuse.
Specific Aims 1) Prevalence/Trends: To determine prevalence rates of chronic prescription opioid use (> 60 days/year) for patients with back pain, arthritis/joint pain, and headaches during 2002-2006 and test for trends over time 2) Initiation: To identify persons with new treatment episodes for these chronic pain conditions, and follow them over time, using Cox proportional hazards models to assess predictors of chronic opioid initiation. 3) Discontinuation: Among all current chronic opioid users, to use Cox proportional hazards models to investigate the predictors of opioid discontinuation . 4) Opioid Abuse/Misuse: Among those who have ever been chronic opioid users, to use logistic and Poisson regression models to investigate predictors of: (a) Onset of diagnosed opioid abuse or dependence, (b) Onset of diagnosed non-opioid substance abuse or dependence, (c) Opioid behaviors suggestive of misuse (multiple pharmacies or prescribers, >365 days opioid supply/yr), (d) Negative health outcomes of opioid use: deaths, hospitalizations, ER visits, overdoses, injuries. 5) Substance Abuse: Among all patients with back pain, arthritis/joint pain, and headaches during 2002- 2006, to use logistic and Poisson regression models to investigate the effect of chronic opioid use on developing: (a) Onset of diagnosed opioid abuse or dependence, (b) Onset of diagnosed non-opioid substance abuse or dependence 6) Risk Stratification: Models derived from 2002-2006 data will be used to formulate risk stratification procedures concerning opioid use which will be validated in the final year (2007) of the study. ? ? ?
Showing the most recent 10 out of 17 publications
