Meta-Regression Models for Historical Data

Suppose that the historical data are available only for the control arm from Kg previous datasets. For the historical data shown in Table 2.2, Kq = 5. Let i/oic 

Comparing (2.12) to (2.3), the models for the historical data and the current data share the common parameters Yo and under the random effects model, while the CV risk effects parameters S and Sq are different in these two models. Similarly, comparing (2.14) to (2.5), the models for the historical data and the current data share the common parameter Yo under the fixed effects model, while the fixed effects parameters 0 and Sq are different in these two models. Thus, the strength of the historical data is borrowed through the common parameters Yo and with different parameters 0 and 0q under the random effects model, and only a single common parameter Yq with different fixed effects parameters 0 and Oq under the fixed effects model provides us with greater flexibility in accommodating different CV risk effects in the current and historical data. 

---