HCR.20.088 – Using clinical prediction models to optimize health care decisions

Route: Health care research, sickness prevention and treatment

Cluster question: 094 How do we improve the quality of health care as much as possible while keeping it affordable?

An important strategy to improve quality of health care while keeping it affordable is to optimize clinical decisions for individual patients. Clinical prediction models (CPMs) are tools to predict benefit and harms of preventive or therapeutic interventions for individual patients, and can be used to target such interventions to those who will benefit, sparing expenses and side effects for those who will not. While CPMs are developed for many fields of medicine, the use of CPMs in clinical practice is limited. We aim to optimize CPM use to increase effectiveness and efficiency of healthcare. We hypothesize that CPM use is hindered by lack of knowledge on: 1) CPM validity in space and time; 2) using CPMs for shared decision making; 3) implementation of CPMs in hospital systems; 4) the impact of using CPMs for clinical decision making. Due to the complexity of this knowledge gap, a large multidisciplinary consortium is needed, including experts in data-driven prediction modeling (experts in statistics, epidemiology and informatics), medical decision making specialists, sociologists, health care professionals, patient organizations, and IT specialists (Electronic Health Record (EHR) developers) from both the public and the private field. We will perform large-scale validity assessment of existing CPMs, using a registry containing over 200 CPMs and publicly available trial databases. In collaboration with the Observational Health Data Sciences and Informatics (OHDSI) program, we will develop techniques for CPM development in EHR data and techniques for continuous validation and updating of CPMs in space and time. We will qualitatively and quantitatively study the process of risk communication and shared decision making based on CPMs. We will study implementation of CPMs in hospital systems and evaluate effectiveness of implementing CPMs in clinical practice. This project will result in a framework for optimal personalized decision making, increasing the effectiveness and efficiency of healthcare.


clinical decision making, Clinical prediction models, EHR data, Healthcare evaluation, Personalized medicine, shared decision making


Organisation Erasmus MC (EMC)
Name Dr. D. (David) van Klaveren
E-mail d.vanklaveren@erasmusmc.nl