An efficient way to do this is to create all the index that are going to belong to the average and then do it in a vectorized way. In your case, you can define each value belonging to the average with intervals of 6 values, an example can be done like this:
N = 15;
A = 1:N; % Simple vector to verify result, substitute by your A with N = 216
idx1 = 1:6:N;
idx2 = 2:6:N;
idx3 = 3:6:N;
Average = mean( [A(idx1);A(idx2);A(idx3)] )'
Average =
2
8
14
