Is the 'but for' test for causation in the law of tort the only appropriate or applicable test?

There exists long standing judicial support for the factual test of causation in the law of tort. For example, in Barnett (1969) the court held that a doctor would not be liable despite negligent treatment, due to the fact that the patient would have died of arsenic poisoning in spite of this negligence. The court has even managed to find novel ways of applying the test in more factually difficult cases, such as Summers, a case involving more than one tortfeasor, where the court applied the test and merely reversed the burden of proof.
That said, Lord Bingham noted in Chester v Afshar (2005) that the factual test of causation may no longer be the exclusive test in the law of tort. This statement is backed by cases like Barker, Sienkiewicz and Fairchild, where it was decided that merely increasing the risk of contracting an illness was sufficient for a causal link to be found. Moreover, as noted in Hotson, the factual causation test is tricky to apply in situations involving a more complex level of uncertainty and percentage risk. Therefore, as Lord Bingham suggests, the but for test is indeed no longer the only applicable or appropriate test forcausation, though it certainly remains the primary test.