Robert B Penfold1, Kelly J Kelleher. 1. Columbus Children's Research Institute, Columbus, Ohio 43205, USA. penfoldr@ccri.net <penfoldr@ccri.net>
Abstract
OBJECTIVE: The purpose of this study was to examine the diffusion of olanzapine to urban and rural children insured by Medicaid in Michigan by identifying prescribing clusters through surveillance of claims records. METHODS: Prescription claims records for all antipsychotic medications for 3,567 children insured by Medicaid in Michigan from 1996 through 1998 were examined through the state Medicaid database. There were 29,069 pediatric prescriptions for antipsychotic medications; 2949 were for olanzapine (576 children, 510 providers). These data were linked to the Area Resource File, Provider Enrollment File, and Rural-Urban Commuting Area codes. Patient and provider locations were geocoded by ZIP code. Mixed logistic regression analysis was performed to determine the probability of a child's being prescribed olanzapine given certain community, patient, and provider characteristics. Spatial clusters were identified through the local Moron's L statistic and empirical Bayes standardized incidence rates. RESULTS: Rural children were more likely than urban children to be prescribed olanzapine (odds ratio [OR], 1.29; P < 0.001). There were significant differences by age and sex, with older children and girls more likely than younger children and boys to be prescribed olanzapine (OR, 1.30 and 1.37, respectively; both, P < 0.001). At the county level, the number of pediatricians per primary care physician reduced the likelihood of a child's being prescribed olanzapine (OR, 0.88; P = 0.039). The effect of the number of available mental health professionals was not significant. The global Moran's L statistic was U indicating moderate clustering of the use of olanzapine. CONCLUSION: Graphic surveillance data may be useful for studying the delivery and use of health cue services. Further research is needed to determine how this method can be used strategically to facilitate or impede the diffusion of new medications.
OBJECTIVE: The purpose of this study was to examine the diffusion of olanzapine to urban and rural children insured by Medicaid in Michigan by identifying prescribing clusters through surveillance of claims records. METHODS: Prescription claims records for all antipsychotic medications for 3,567 children insured by Medicaid in Michigan from 1996 through 1998 were examined through the state Medicaid database. There were 29,069 pediatric prescriptions for antipsychotic medications; 2949 were for olanzapine (576 children, 510 providers). These data were linked to the Area Resource File, Provider Enrollment File, and Rural-Urban Commuting Area codes. Patient and provider locations were geocoded by ZIP code. Mixed logistic regression analysis was performed to determine the probability of a child's being prescribed olanzapine given certain community, patient, and provider characteristics. Spatial clusters were identified through the local Moron's L statistic and empirical Bayes standardized incidence rates. RESULTS: Rural children were more likely than urban children to be prescribed olanzapine (odds ratio [OR], 1.29; P < 0.001). There were significant differences by age and sex, with older children and girls more likely than younger children and boys to be prescribed olanzapine (OR, 1.30 and 1.37, respectively; both, P < 0.001). At the county level, the number of pediatricians per primary care physician reduced the likelihood of a child's being prescribed olanzapine (OR, 0.88; P = 0.039). The effect of the number of available mental health professionals was not significant. The global Moran's L statistic was U indicating moderate clustering of the use of olanzapine. CONCLUSION: Graphic surveillance data may be useful for studying the delivery and use of health cue services. Further research is needed to determine how this method can be used strategically to facilitate or impede the diffusion of new medications.
Authors: Robert B Penfold; James F Burgess; Austin F Lee; Mingfei Li; Christopher J Miller; Marjorie Nealon Seibert; Todd P Semla; David C Mohr; Lewis E Kazis; Mark S Bauer Journal: Health Serv Res Date: 2016-12-22 Impact factor: 3.402