That paper is extremely weak. Look at the confidence intervals surrounding the hazard ratios. There is almost no effect for paracetamol and the difference between paracetamol and ibuprofen is negligible. If this is the evidence, I’m not concerned.