ABSTRACT
Nuanced Interpretations of Data . . . . . . . . . . . . . . . . . . . . 326 10.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331
10.4.1 Need for Better Primary Studies and Reporting of Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332
10.4.2 Need to Develop Methods to Optimally Engage Stakeholders and Patients . . . . . . . . . . . . . . . . . . . . . . . . . 333
10.4.3 Need to Incorporate Information from Different Designs and to Combine Individual and Aggregate Data . . . . . . . . 333
10.4.4 Need to Modernize Review Methods . . . . . . . . . . . . . . . . 334 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
ABSTRACT Systematic review and meta-analysis are the fundamental tools of evidence-based healthcare. Combining information from multiple studies can provide insights that individual studies cannot offer. In this chapter we explain these methods and show how they are applied by AHRQ (a US-based healthcare agency) in its comparative effectiveness research program. We illustrate the challenges of conducting comparative effectiveness reviews using study level data in three examples of reports produced by the AHRQ evidence-based practice centers. In the second half of this chapter, we provide a detailed discussion of the statistical methods used in individual patient data meta-analysis. The approach to combine patient level data has the ability to answer questions that are directly applicable to individual patients, whereas study level meta-analysis can only offer conclusions for a population. We anticipate that individual patient data will likely be more readily available for future meta-analyses. The increasing demand for comparative effectiveness reviews will need more efficient methods to produce them. We offer some suggestions to modernize systematic review and meta-analyses methods.
