Patient monitoring is routinely used to detect disease aggravation in many chronic conditions. We propose an adaptive scheduling strategy based on dynamic individual risk predictions that can improve the efficiency of monitoring programs that incorporate multiple longitudinal measurements and competing events. It is motivated by stable chronic heart failure (CHF) patients who are periodically seen to assess the risk of disease aggravation based on multiple patient characteristics and circulating marker protein levels such as NT-proBNP and troponin. We assess the performance of the adaptive strategy versus fixed schedule alternatives using a simulation study based on the Bio-SHiFT study, a cohort of stable CHF patients.