Can Machine Learning from Real-World Data Support Drug Treatment Decisions? A Prediction Modeling Case for Direct Oral Anticoagulants

Abstract

Background: Decision making for the “best” treatment is particularly challenging in situations in which individual patient response to drugs can largely differ from average treatment effects. By estimating individual treatment effects (ITEs), we aimed to demonstrate how strokes, major bleeding events, and a composite of both could be reduced by model-assisted recommendations for a particular direct oral anticoagulant (DOAC). Methods: In German claims data for the calendar years 2014–2018, we selected 29 901 new users of the DOACs rivaroxaban and apixaban. Random forests considered binary events within 1 y to estimate ITEs under each DOAC according to the X-learner algorithm with 29 potential effect modifiers; treatment recommendations were based on these estimated ITEs. Model performance was evaluated by the c-for-benefit statistics, absolute risk reduction (ARR), and absolute risk difference (ARD) by trial emulation. Results: A significant proportion of patients would be recommended a different treatment option than they actually received. The stroke model significantly discriminated patients for higher benefit and thus indicated improved decisions by reduced outcomes (c-for-benefit: 0.56; 95% confidence interval [0.52; 0.60]). In the group with apixaban recommendation, the model also improved the composite endpoint (ARR: 1.69 % [0.39; 2.97]). In trial emulations, model-assisted recommendations significantly reduced the composite event rate (ARD: −0.78 % [−1.40; −0.03]). Conclusions: If prescribers are undecided about the potential benefits of different treatment options, ITEs can support decision making, especially if evidence is inconclusive, risk-benefit profiles of therapeutic alternatives differ significantly, and the patients’ complexity deviates from “typical” study populations. In the exemplary case for DOACs and potentially in other situations, the significant impact could also become practically relevant if recommendations were available in an automated way as part of decision making. Highlights It was possible to calculate individual treatment effects (ITEs) from routine claims data for rivaroxaban and apixaban, and the characteristics between the groups with recommendation for one or the other option differed significantly. ITEs resulted in recommendations that were significantly superior to usual (observed) treatment allocations in terms of absolute risk reduction, both separately for stroke and in the composite endpoint of stroke and major bleeding. When similar patients from routine data were selected (precision cohorts) for patients with a strong recommendation for one option or the other, those similar patients under the respective recommendation showed a significantly better prognosis compared with the alternative option. Many steps may still be needed on the way to clinical practice, but the principle of decision support developed from routine data may point the way toward future decision-making processes.

Establishing the worth of deprescribing inappropriate medications: are we there yet?

“…Machine learning applied to largescale clinical trial or pharmacovigilance data are identifying patient phenotypes at greater risk of medication-specific harm to whom deprescribing interventions could be targeted. 21,22 Fourth, sensitive medication-related quality of life measures relevant to both medication class-specific effects and individual goals of care may be better able to detect subtle but important patient-centred outcomes. 23 Finally, n-of-1 trials, in which patients act as their own control during randomised cycles of exposure to a drug or placebo, may provide more nuanced assessment of Perspectives deprescribing effects in individuals.…”

“…Third, where evidence reveals factors predicting individuals more likely to benefit or be harmed by withdrawing problematic medications, such as anticonvulsant 19 and antihypertensive drugs, 20 these should be integrated into guidelines and decision support systems. Machine learning applied to large‐scale clinical trial or pharmacovigilance data are identifying patient phenotypes at greater risk of medication‐specific harm to whom deprescribing interventions could be targeted 21,22 …”

“…Machine learning applied to large‐scale clinical trial or pharmacovigilance data are identifying patient phenotypes at greater risk of medication‐specific harm to whom deprescribing interventions could be targeted. 21 , 22 …”

The role of INTERCheckWEB digital innovation in supporting polytherapy management

“…Any methods of facilitating guidance on managing patients with polypharmacy, through the development and application of ‘risk prediction tools’ for quantifying the risk of adverse drug reactions, acquire a strategic relevance 20 , 21 . In this view, clinicians may benefit from considering evidence-based recommendations of drug use to preserve patient safety, worldwide.…”

A Practical Statin Recommendation System Based on Real-World Data to Improve LDL-C Management in Secondary Prevention

“…Recently, machine learning has been studied in precision medication to help decision making in an individual patient. [9][10][11] Although results from large population-based studies provide average treatment effects, physicians are unsure to make the best strategy for a single patient due to potential individual heterogeneity. Machine learning can be used to approach complex clinical situations and train predictive models based on single individuals.…”