Definitions of terms within "rough" view are as follows. A test that is useless will not change the probability at all. An "almost useless" test will increase a 1% probability to no more than 2% and decrease a 10% probability to no less than 5%. A "fair" test will increase a 1% probability to 2-10% and decrease a 10% probability to 1-5%. A "good" test will increase a 1% probability to 10-30% and decrease a 10% probability to 0.1-1%. An "excellent" test will increase a 1% probability to >50% and decrease a 10% probability to <0.1%. A test that is "unclear" was not precise enough in the original study to make any firm conclusions as to its accuracy.