Bell's inequality predicts a value for the correlations of entangled pairs must be greater than a certain value, if there are predetermined hidden variables encoded in the particles at the time of emission. If the correlations dot exceed this value then then it predicts there can be be no predetermined hidden variables and that the outcome of measurements is determined at the time the measurement is made. This is condition is called a 'violation' of Bell's inequalities and it implies non local interactions (i.e .action at a distance exceeding the speed of light) between the particles.
Bell's paper introducing his inequalities in 1964 was a response (nearly 30 years later!) to a thought experiment called the EPR paradox, that was thought up (nearly 30 years earlier) by Einstein, Podolsky and Rosen in 1935, to demonstrate that Quantum Mechanics must be incomplete, because it appears to imply non local superluminal interactions, which of course Einstein would be very much against because his Relativity theory insists nothing, not even communication, can exceed the speed of light. Unfortunately for Einstein, later experiments confirmed that Bell's inequalities are violated and that non local interactions must be taking place between entangles pairs.
Does this invalidate Relativity? No. this is because the non local interactions occurring between distant particles at the quantum level, can not be used in any way by physical observes to send communication signals between themselves faster than light, nor can any physical interactions between macroscopic objects, or changes in physical fields, occur at greater than the speed of light.
Now to focus on the issue of determinism, the orientation of polarisation (or other observables) of the emitted entangled pairs is purely random. Bell's inequalities addresses the issue of whether these observables are fixed at the time of emission (pre-determined) or at the time of measurement. Bell predicted that they are determined at the time of measurement and subsequent experiments confirmed that prediction. In other words If the outcome of experiments find correlations that do not violate Bell's inequalities, then that concurs with the point of view that the correlations' are deterministic in the sense that they are determined at the time of emission, but experiments show that is not the case.
There are of course many interpretations of Quantum Mechanics, some such as the 'many worlds' interpretation that says every outcome occurs simultaneously and we just see one possible outcome and there is also the even more deterministic 'pilot wave' interpretation. Bell himself never really mentioned determinism in his paper, but focused on locality and hidden variables. If you consider hidden variables to be deterministic, then a non-violation of his inequalities would be a proof of determinism, but that has not been observed in reality.