The question assumes interstellar radio communications is possible in the first place. We actually don't know this is true, because we've never successfully contacted any artificial EMR-producing device beyond our own deep space probes in the Kuiper Belt, in the outskirts of our own Solar System (but still only about 6 light-hours from the Sun).
The big problem we have is the inverse-square law. Most of the radio transmissions we generate as a race that leave our neighborhood are more or less omnidirectional. That means that the power of the transmission, as received by an antenna, decreases on the square of distance, because all other things being equal your antenna is intercepting some fraction of the surface area of a sphere of the transmitter's broadcast, so at a constant carrier wave power produced by an antenna as essentially a point source, the surface area that energy is spread across is roughly (4/3)πr2, so as radius increases, the surface area increases on the square of radius, and therefore your antenna area as a fraction of surface area decreases on the square.
Currently, the most powerful single artificial radio signal generator we have is the Taldom Transmitter in Russia. It broadcasts on two frequencies, the higher one (261kHz) is generated at 2500kW or 2.5MW. That sounds like a lot, and it is, but the very nearest star is 4.25 light years away from us.
Radio waves can be measured in Janskys (it's a non-SI unit but based on metric measurements so it can be easily converted to the SI watt). A Jansky is a unit of wave power density equal to 10-26W/m2/Hz. The unit is common in radio astronomy because astronomical objects like stars involve massive amounts of EM flux acting across very long distances. Most astronomical objects outside our solar system have EM flux in the range of 1 to 100 Janskys.
Given the transmitter's frequency and strength, the radio flux density of the Taldom Transmitter as detected by a ship orbiting Proxima Centauri would be on the order of:
$$2500000 / \dfrac{4\pi}{3}(4.25 * 9.461^{15})^2 / 261000 * 10^{26} = 1.414 * 10 ^{-7}\text{Jy}$$
That's 141.4 nanoJanskys. And this back-of-the-envelope calculation doesn't account for free-space path loss and other phenomena involved in interstellar EMR transmission. Basically, if you're broadcasting for anyone to hear, nobody can hear you.
We could improve on that, dramaticaly, by making the signal directional. Ideally, if the Taldom Transmitter's power were directionalized with a parabolic antenna, say one similar to Arecibo (152m radius, 72583m2 area) that send the radio power laser-like out into space toward Proxima, the signal strength would be:
$$2500000/72583/261000 * 10^{26} = 13.197 * 10^{21} \text{Jy}$$
That's about a quintillion times the normal radio emissions of anything else in our neighborhood, so assuming the antenna was exactly spotlighting the receiver, your transmission would blow out a dime-store transistor radio. But, all you'd need to block it is one not-very-large asteroid along the transmission path (or a misalignment of the transmitter antenna by picoseconds of arc; trying to hit your target planet with a radio transmission perfectly focused at 305 meters width of beam, in the grand scheme of things, is like trying to hit the bullseye on a dartboard, mounted on a G-simulation centrifuge at top speed, from 3 miles away).
So, the answer is, you can very easily prevent the transmission of radio information between two communicating entities. An omni-directional "beacon" type signal, you wouldn't even have to block. A more directional signal could be physically blocked or misdirected with a solar sail or similar device somewhere directly in line between transmitter and receiver.