It does. It is just that since (if) the slit separation is substantially greater than the slit widths this factor dominates. You can however do a more precise calculation which takes both effects into account.
[edit: PS] My reasoning may not quite be right w.r.t. relative widths but the rest is correct.[ end edit]
I agree with jambaugh. Another way of saying it is that, for a single slit difraction patern, you assume that the width of the slit is nonnegligible, whereas, for a double slit pattern, you assume that the slits are infinitesimals, like ideal line sources with zero width.