How to combine linear mixed model results across different test subjects?

Suppose I have three test subjects and for each I have collected data and formulated a linear mixed-effects model using fitlme. (Assume the effects, both fixed and random, are the same for each.)

If LME is the resulting model for a particular individual, the command anova(LME), gives me a p-value for each fixed effect, indicating its significance in determining the response outcome.

How can I combine the three anova analyses to test whether the fixed effect in general plays a role in determining the response?