Key Points
• Source: medRxiv
• Conclusion/Relevance: “We conclude that the highly distributed mode of causation in chronic migraine necessitates high-dimensional modelling for optimal management. Machine prescription should be considered an essential clinical decision-support tool in the future management of chronic migraine.”
• In the current study, investigators analyzed 1446 patients with chronic migraine. They estimated individual treatment effects spanning 10 classes of preventives to determine potential causes for response.
• They found that the machine prescription policy results in an estimated 35% reduction in time-to-response vs expert guidelines, with no cost increase per patient.
• The UK and Norwegian authors suggest that their study is largest and best phenotyped to date. They found treatment heterogeneity was comprehensively predictable based on high-dimensional causal modelling of routinely collected clinical data.
• The authors wrote, “This finding supports a complex, distributed underlying mode of causation in chronic migraine, and suggests that neither the pursuit of a unitary causal mechanism, nor the evaluation of treatment effects within conventional randomised controlled trials is likely to be productive. Rather, deeper characterisation of patient heterogeneity is likely to be needed, through modelling richer additional features, such as imaging, physiological and genetic data at larger data scales, illuminating the wide causal field of factors that clearly underpins this complex disorder.”
• Importantly, the investigators found that current treatment policy guidelines likely result in the same time-to-response as that of chance.
• Limitations of the current study include some patients not meeting strict diagnostic criteria, although evaluated by expert headache specialist to have chronic migraine, Additionally, allocation bias could have been an issue due to the observational design.