Aims: To determine the cost-effectiveness of alternative models of practice nurse involvement in the management of type 2 diabetes within the primary care setting. Methods: Linked routinely collected clinical data and resource use (general practitioner visits, hospital services and pharmaceuticals) were used to undertake a risk-adjusted cost-effectiveness analysis of alternative models of care for the management of diabetes patients. These models were based on the reported level of involvement of practice nurses in the provision of clinical-based activities. Potential confounders were controlled for by using propensity score-weighted regression analyses. The impact of alternative models of care on outcomes and costs was measured and incremental cost-effectiveness estimated. The uncertainty around the estimates of cost-effectiveness was illustrated through bootstrapping. Results: Although the difference in total cost between two models of care was not statistically significant, the high-level model was associated with better outcomes (larger mean reductions in HbA1c). The upper 95% confidence intervals showed that the incremental cost per 1% decrease in HbA1c is only $454, and per one additional patient to achieve an HbA1c value of less than 53 mmol/mol (7.0%) is $323. Further analyses showed little uncertainty surrounding the decision to adopt the high-level model. Conclusions: The results provide a strong indication that the high-level model is a cost-effective way of managing diabetes patients. Our findings highlight the need for effective incentives to encourage general practices to better integrate practice nurses in the provision of clinical services.