Best practices
Promote accrual through leadership best practices and organizational development. This may include establishing a “culture of commitment” to clinical trials from the highest levels at a site as part of standard of care, including multidisciplinary teams that prioritize clinical trials. Implement site and clinical trialist performance standards that qualify clinical investigators based in part on their accrual performance. Use available site data, including screening logs, to verify that patient populations are available for trials the site is considering, and to identify patient populations at the local level who lack available trials. Promote leadership/ownership of investigator-initiated trials. Use formal quality improvement techniques to increase the efficiency of opening and conducting trials. Use a clinical trials management system as a tool to track the various aspects of managing protocols and empower a Steering Committee or core team to utilize it to evaluate site progress. Close trials that fail to accrue at a reasonable pace, with allowances for variable rates for less common tumor types or more rare disease settings. Use the NCI's Central IRB (CIRB) for NCI Cooperative Group trials to shorten IRB turnaround times and reduce workload of local IRBs, particularly with the new CIRB Independent Model.
|
Site leadership, organization, and operations
Study both successfully accruing and poorly accruing sites for lessons learned regarding their organizational and leadership strategies, as well as their infrastructure, staffing, and trials portfolio. Test new organizational models and evaluate their impact on accrual. Models are especially needed for private-practice trials unsupported by an academic institution. Identify, implement, and then evaluate leadership models effective in other fields, such as commercial enterprise, that may also be applicable to improving accrual to clinical trials. Consider engaging anthropologists, sociologists, and operations engineers in developing these leadership models.
|
