Any edge deflects photons into areas with higher and lower density. To measure such a result different measurement instruments can be used. Beside a screen one may use a photodiode which is moved and the outcoming current shows the same intensity distribution like the pattern on a screen.
The detection of photons direct behind the slit destroys the interaction between the photons and the edge(s). Since a photon-photon interaction is very improbable you could use electrons, shooting them parallel to the wall with the slits. It will interfere the fringes. But why?
We know that the electrons we use for detection, interact with the photons after the edges. With the same certainty, we should assume that the photons and the surface electrons of the edges interacting.
Let's imagine an experiment. We shoot three electron beam and set up our photon beam perpendicular to it. We should get a result with an intensity distribution similar to the double slit (not the same, because only a small amount of photons meets the electrons).
Now, if you try to find out, which way the photon was going between the electron beams and you do this with an additional electron beam, clearly you’ll disturb the intensity distribution. Simply, we not have the sensitive enough measurement instrument.