OBJECTIVES: This project presents an analysis of IT processes on variations of prescribing patterns for patients diagnosed with diabetes type II, following the first study on electronic billing and its association with diabetic drug prescribing  (Huttin/Wong,2010). METHODS: A sample of 610 patients is extracted from the CDC physician survey.IT processes include electronic medical records (EMR), with or without patient demographics, computerized orders for Rx, tests, lab results, notes from nurses and physicians, public health reporting. Several  hierarchical clustering methods are tested to identify various stages of IT processes and the impact of IT  is analyzed with non parametric tests (analysis of variance) on prescribing patterns. RESULTS: Two HB clustering methods (average method and ward)  identify three clusters representing different stages of IT processes: physicians using no IT at all (80.76%) and two levels of IT operations in practices (cluster 2: 17.33 %; cluster 3: 1.91%). The dendogram with the AL method presents clearer separation than the dendogram with the ward method. Variations in drug prescribing is significant  between clusters, using the scoring savage test: -21 for cluster 1, 16.85 for cluster 2, 4.4 for cluster 3 (P value 0.01). The analysis on new drugs does not show different prescribing patterns;  however, the number of injectables (insulin) per patient is significantly higher in cluster 2 than 1 (0.51 verus 0.39) . Different patterns of IT processes are also identified within cluster 2 and other clustering methods among grouping and similarity computations (e.g. Shusaku et als,2004)are tested to analyze the propagation of IT processes among the practices of this dataset (generalisation tested with a similarity matrix). CONCLUSIONS: This project can be used for analysis and management of IT processes inside clinical systems and control for their effects on physician prescribing behaviours.Results confirm that in addition to ebilling, different patterns of IT processes have an impact on treatment regimens (especially affecting insulin or insulin/OAD combinations). This can complement Koro et als study (2004).