In diabetes patients with uncontrolled glucose, BP, and hyperlipidemia, the question of prioritization arises: what should be treated first? To address this issue, we created a simulated population of 10,000 diabetes patients modeled to match the distribution of a real patient population in Minnesota. We then applied a sophisticated computational model to simulate and compare the clinical results and costs of several distinct treatment strategies applied for up to 6 visits: (a) treat parameter furthest from goal; (b) treat hyperlipidemia to goal first; (c) treat hypertension to goal first; (d) treat hyperglycemia to goal first; and (e) treat the condition with the greatest contribution to macrovascular risk (using the UKPDS Risk Engine estimates). For comparison purposes, we also simulated (f) continuation of current treatment without additional intensification as a lower bound and (g) aggressive treatment, under which intensification for all three conditions takes place at every visit, as an upper bound. Strategies (b), (c), and (d) were inferior to other strategies. Strategy (e) outperforms strategy (a) by reducing the CHD events from 1436 to 1299 and the stroke events from 1281 to 1236 after 6 visits. The costs of (e) and (a) were not significantly different, but both these strategies had much lower costs than strategy (g), which is the strategy currently recommended by many experts and which had only marginally better clinical outcomes. This modeling approach can estimate comparative clinical and cost effectiveness of treatment strategies for both populations and individual patients, since the simulation platform takes individual patient characteristics (such as A1c, SBP, LDL, oral medications and insulin) into account.