Yeah, I'd really like to see that statistic corrected for long-term prescriptions. Or better yet, controlled for diagnosis.
Following that link, we see that patients issued 30 days of opioids via multiple prescriptions have a 30% chance of one year use, but patients issued 30 days up front have a 45% chance.
Interpreting that as evidence for shorter prescription windows looks like unproven causation to me. There's presumably a reason some patients get 30 days as their initial dose, and that's not being controlled for.
The study suggests that since only 1% of patients got strong and long-lasting opioids, prescriptions for chronic conditions must be rare, but that seems wrong twice over. First, as noted in the article, 10% got tramadol at some dose, and may well be chronic patients. Second, recurring health issues are overlooked; a patient getting 30 days of opioids up front may not have a chronic condition, but might well have a recurring issue that leaves them needing subsequent scripts.
More broadly, we seem to be attempting to control opioids on the supply side in particularly foolish ways. It's already known that major problems include over-prescribing doctors and dosage cycles that produce withdrawal, but attempts to treat the problem focus on restricting well-meaning prescribers.