The article and abstract mention self-reporting. They probably haven't been able to objectively measure pain, sleep, emotional problems. The abstract says the change in opoid use are not statistically significant.
There's already a number of study with conflicting results on Canabis effect on pain. Some show an effect on pain reduction. Some show no effect when controlling for the placebo effect.
If that study doesn't have better methodology than previous ones, no control for placebo, isn't able to link the effect to a specific molecule, then it probably won't help settle the question.