Ongoing efforts to scientifically evaluate the methods used to conduct systematic reviews are critical for ensuring the validity and credibility of current and future systematic reviews. As the largest producers of systematic reviews in North America, AHRQ's Evidence-based Practice Centers are also leaders in advancing the methodology of conducting systematic reviews.
Methods development and refinement is a natural outgrowth of many of our projects, and many important methodological publications have resulted from this work. KP EPC investigators participate in workgroups of AHRQ's Effective Health Care Program to advance the methods of conducting systematic reviews. We have collaborated in workgroups on addressing health equity in systematic reviews and evidence-based clinical practice guidelines, data visualization for complex data, utilizing AI in systematic reviews, synthesis of qualitative studies, using existing systematic reviews, multi-component interventions, risk stratified analyses and recommendations, quality improvement, as well as advancing risk of bias assessment and quantitative analysis approaches.
Our researchers have also been deeply involved in methods and procedural development for the U.S. Preventive Services Task Force (USPSTF) for almost 20 years. We provide ongoing support for four USPSTF workgroups that focus on methodological issues or population groups, including workgroups on methods, topic prioritization, child and maternal health, and dissemination and implementation.
KP EPC researchers have continued to help the USPSTF refine its approaches to updating reviews, incorporating decision analyses with systematic reviews, and adapting the USPSTF methods to fairly address the needs of specific populations. In addition, we have helped create and update the USPSTF Procedure Manual in collaboration with USPSTF members and AHRQ, and have led efforts to articulate, implement, and refine methods for the USPSTF across a wide range of methodological areas, including development of a health equity framework, algorithmic bias in risk-prediction models, non-randomized studies of interventions, inclusion of comparative effectiveness studies, and intensity of complex behavioral interventions.