A scoping review of causal methods enabling predictions under hypothetical interventions

Abstract

Background The methods with which prediction models are usually developed mean that neither the parameters nor the predictions should be interpreted causally. For many applications, this is perfectly acceptable. However, when prediction models are used to support decision making, there is often a need for predicting outcomes under hypothetical interventions. Aims We aimed to identify published methods for developing and validating prediction models that enable risk estimation of outcomes under hypothetical interventions, utilizing causal inference. We aimed to identify the main methodological approaches, their underlying assumptions, targeted estimands, and potential pitfalls and challenges with using the method. Finally, we aimed to highlight unresolved methodological challenges. Methods We systematically reviewed literature published by December 2019, considering papers in the health domain that used causal considerations to enable prediction models to be used for predictions under hypothetical interventions. We included both methodologies proposed in statistical/machine learning literature and methodologies used in applied studies. Results We identified 4919 papers through database searches and a further 115 papers through manual searches. Of these, 87 papers were retained for full-text screening, of which 13 were selected for inclusion. We found papers from both the statistical and the machine learning literature. Most of the identified methods for causal inference from observational data were based on marginal structural models and g-estimation. Conclusions There exist two broad methodological approaches for allowing prediction under hypothetical intervention into clinical prediction models: (1) enriching prediction models derived from observational studies with estimated causal effects from clinical trials and meta-analyses and (2) estimating prediction models and causal effects directly from observational data. These methods require extending to dynamic treatment regimes, and consideration of multiple interventions to operationalise a clinical decision support system. Techniques for validating ‘causal prediction models’ are still in their infancy.

Using clinical prediction models to personalise lifestyle interventions for cardiovascular disease prevention: A systematic literature review

“…A CPM may be combined with relative risk reduction estimates from RCTs, as described by Harrell and Lazzeroni (2017) . To derive causal CPMs based on observational data, other techniques have been developed ( Lin et al, 2021 , Shalit et al, 2017 ). Additionally, suggested actions may be based on combining non-causal CPMs and decision-curve analyses ( Vickers and Elkin, 2006 ).…”

A tutorial on individualized treatment effect prediction from randomized trials with a binary endpoint

“…Such topics require careful attention to the exchangeability assumption, which is no longer fulfilled by the study design and needs further assumptions and careful modeling with respect to all possible confounders. A recent scoping review provides an overview of the literature with respect to methods for causal prediction that extend to observational data 65 . Also, where we have focused on intention to treat estimates of point exposure treatment, different settings and questions require further thought on the relevant definition of the estimand 66 .…”

“…Hence, if we want to walk down the avenue of individualized treatment effect modeling, we will either have to design trials with this purpose in mind, or have to find more creative ways to amplify our data. This could include the analysis of individual patient data from multiple randomized trials, or even the use of nonrandomized studies for the estimation of outcome risk under a control condition 65 . Besides clear opportunities, such approaches also bring about many new challenges.…”

Continuing to Advance Epidemiology

“…The increasing availability of data alone will not help improve population health if we only observe them without interpreting the underlying reality (10). New, innovative concepts are being developed, mainly in the intersection between predictive and causal frameworks (11)(12)(13) that will help to improve our understanding and provide advances.…”