Predictors of long-acting opioid use and oral versus transdermal route among older medicaid beneficiaries

Abstract

The availability of long-acting (LA) opioid medications represents an advance in the treatment of persistent pain, raising the question of what factors predict receipt of this newer treatment approach. This study sought to determine which factors predict use of LA opioid medications and which factors predict use of transdermal versus oral options in an older population. A retrospective analysis was conducted of Medicaid claims from May 1999 to April 2002 for older (aged >60 years) Kansas Medicaid beneficiaries receiving opioid medications equivalent to >or=600 mg of morphine per month. Demographic data, diagnoses, and medication use were derived from Medicaid claims data provided by the Kansas Department of Social and Rehabilitative Services to Dr. Shireman for a larger project. Bivariate analyses and multivariable models were used to determine which clinical and demographic factors predicted LA opioid use in general and which factors were associated with transdermal fentanyl (TDF) use in particular. The study population was mostly women, and more than one third were aged >85 years. Of the 766 claims, those patients taking LA opioids (n = 352) received oral sustained-release morphine, oral controlled-release oxycodone, or the TDF patch, with or without concomitant short-acting (SA) opioids. All others received SA opioids only (n = 414). In multivariate analysis, LA opioid use was more common among nursing home residents (adjusted odds ratio [AOR] = 1.45; 95% CI, 1.04-2.02) and persons under hospice care (AOR = 2.10; 95% CI, 1.19-3.73), with malignancy also bordering on a statistically significant association (AOR = 1.63; 95% CI, 1.00-2.66). Age, sex, race/ethnicity, and urban/nonurban residence were not associated with LA opioid use. Among those taking LA opioids, TDF use was more likely with older age (AOR = 2.89; 95% CI, 1.44-5.81 for aged >or=85 years vs aged 60-75 years), nursing home residence (AOR = 2.51; 95% CI, 1.46-4.49), dementia (AOR = 2.77; 95% CI, 1.03-7.48), malignancy (AOR = 3.03; 95% CI, 1.25-7.34), and nonwhite race (AOR = 2.72; 95% CI, 1.18-6.28), and less likely with chronic obstructive pulmonary disease (AOR = 0.47; 95% CI, 0.26-0.88), in adjusted analysis. Demographic and clinical predictors of TDF use differed from predictors of any LA opioid use among these older Medicaid beneficiaries. Qualitative research methods are needed to better understand what factors influence selection of opioid medications for older patients in nursing homes and community settings.