Generalised decision curve analysis for explicit comparison of treatment effects

Abstract

RationaleDecision curve analysis (DCA) helps integrate prediction models with treatment assessments to guide personalised therapeutic choices among multiple treatment options. However, the current versions of DCA do not explicitly model treatment effects in the analysis but implicitly or holistically assess therapeutic benefits and harms. In addition, the existing DCA cannot allow the comparison of multiple treatments using a standard metric.Aims and ObjectivesTo develop a generalised version of DCA (gDCA) by decomposing holistically assessed net benefits and harms into patient preferences versus empirical evidence (as obtained in the trials, meta‐analyses of clinical studies, etc.) to allow individualised comparison of single or multiple treatments using a common metric.MethodsWe reformulated DCA by (1) decomposing holistic, implicit utilities into specific utilities related to treatment effects and patient's relative values (RV) about disease outcomes versus treatment harms, (2) explicitly modelling each treatment effect at the level of probabilities and/or utilities (outcomes) in a decision tree, and (3) avoiding scaling effects employed in the original DCA to enable comparison of treatment effects against the common metrics. We used data from a published network meta‐analysis of randomised trials to inform the use of statin treatment according to Framingham Risk Model.ResultsWe illustrate the analysis by modelling the effects of three statins in the primary prevention of cardiovascular disease. We performed simultaneous comparisons against standard metrics (RV) for all treatments. We examined for which RV values, a predictive model for guiding personalised treatment, outperformed the strategies of treating everyone or treating no one. We found that the magnitude of benefits (efficacy) seems more important than the simple ratio of efficacy/harms.ConclusionWe describe gDCA for evaluating single or multiple treatments to help tailor therapy toward individual risk characteristics. gDCA further helps integrate the principles of evidence‐based medicine with decision analysis.

Decision curve analysis based on summary data

“…This is because predictive models do not assess the consequences (benefits and harms) of competing treatment alternatives. 4 The latter is possible if the predictive models are integrated within a decision-theoretical framework such as DCA. [1][2][3][4] In turn, this allows the generation of risk-adapted personalized treatment choices to select the most optimal management strategy.…”

“…4 The latter is possible if the predictive models are integrated within a decision-theoretical framework such as DCA. [1][2][3][4] In turn, this allows the generation of risk-adapted personalized treatment choices to select the most optimal management strategy. [1][2][3][4] In other words, the predictive model can separate patients into risk groups, some patients are at lower and some at higher risk for the event of interest.…”

“…[1][2][3][4] In turn, this allows the generation of risk-adapted personalized treatment choices to select the most optimal management strategy. [1][2][3][4] In other words, the predictive model can separate patients into risk groups, some patients are at lower and some at higher risk for the event of interest. 8,9,[11][12][13][14][15] Because treatment effects increase with baseline risk, the patients at higher risk of the event will benefit more than those at lower risk.…”

“…8,9,[11][12][13][14][15] DCA facilitates individualization of treatments because only those patients whose risk of clinical outcome exceeds the given threshold receive treatment. 4,9 This ensures that the 'average' effects from the metaanalyses are translated into more personalized treatment. 4,9 Figure 1 shows a conceptual relationship between metaanalyses, predictive models and DCA.…”

“…Decision curve analysis (DCA), a widely used method for evaluating diagnostic tests and predictive models, has thus far required individual patient data (IPD). [1][2][3][4] Although IPD remains optimal for the development of predictive models, they are relatively rarely done. 5 Even when investigators develop models based on IPD, they are often not subjected to decision analytical assessments such as DCA.…”