The risk at 1/10000 is on the high side for those who assume a linear model vs a linear quadratic model when extrapolating known data for instance hiroshima or chernobyl exposure. Let me put the risk in a simpler way. About 1,000 cases of cancer per year are estimated to be caused in the US a year from the total population exposure to diagnostic x rays.
One must exercise some prudence. So for example, i dont allow routine dental xrays everytime I see a dentist. On the other hand a test to diagnose, treat, or stage cancer or the high exposure during cardiac cath is well worth the risk. If a pregnacy is suspected then we get a pregnancy test before exposure.
One must also be aware of background radation. If you live in higher altitudes, travel a lot on planes, live in the radon belt or an adobe hut you get a lot more radiation from the environment. Then theres the occasional solar flare and of course your risk of skin cancer when you get a sunburn from ultraviolet radiation.