Yapay zekâ opioid reçetelerini daha isabetli kılabilir mi?

Pain after the same dental operation can vary greatly between patients, yet opioid prescriptions are often standardised before clinicians know how each person will recover. A Harvard School of Dental Medicine paper considers whether machine learning could support more individual decisions. Models could combine clinical, behavioural, biological and social information to estimate who may need stronger pain relief, who faces a higher risk of prolonged use or misuse, and who would benefit from closer monitoring. This could reduce “just-in-case” prescribing: studies cited by the authors found that only 7 percent of patients needed opioids after routine wisdom-tooth surgery, while more than half of prescribed tablets after dental procedures went unused. However, the goal is not simply to minimise prescriptions, because some patients genuinely need additional relief. The authors also stress that most existing studies are retrospective, external validation is limited, and results may differ across populations. AI should therefore inform clinical judgement rather than replace it.