Dynamic predictive probabilities to monitor rapid cystic fibrosis disease progression

Abstract

Cystic fibrosis (CF) is a progressive, genetic disease characterized by frequent, prolonged drops in lung function. Accurately predicting rapid underlying lung‐function decline is essential for clinical decision support and timely intervention. Determining whether an individual is experiencing a period of rapid decline is complicated due to its heterogeneous timing and extent, and error component of the measured lung function. We construct individualized predictive probabilities for “nowcasting” rapid decline. We assume each patient's true longitudinal lung function, S(t), follows a nonlinear, nonstationary stochastic process, and accommodate between‐patient heterogeneity through random effects. Corresponding lung‐function decline at time t is defined as the rate of change, S′(t). We predict S′(t) conditional on observed covariate and measurement history by modeling a measured lung function as a noisy version of S(t). The method is applied to data on 30 879 US CF Registry patients. Results are contrasted with a currently employed decision rule using single‐center data on 212 individuals. Rapid decline is identified earlier using predictive probabilities than the center's currently employed decision rule (mean difference: 0.65 years; 95% confidence interval (CI): 0.41, 0.89). We constructed a bootstrapping algorithm to obtain CIs for predictive probabilities. We illustrate real‐time implementation with R Shiny. Predictive accuracy is investigated using empirical simulations, which suggest this approach more accurately detects peak decline, compared with a uniform threshold of rapid decline. Median area under the ROC curve estimates (Q1‐Q3) were 0.817 (0.814‐0.822) and 0.745 (0.741‐0.747), respectively, implying reasonable accuracy for both. This article demonstrates how individualized rate of change estimates can be coupled with probabilistic predictive inference and implementation for a useful medical‐monitoring approach.

Cystic Fibrosis Point of Personalized Detection (CFPOPD): An Interactive Web Application

“…The source of patient data used during CFPOPD development and the algorithm’s development and validation has been described in detail elsewhere [ 8 ]. Briefly, we obtained data for 30,879 patients from the US CFFPR from 2003 to 2015 to train and validate our algorithm.…”

“…CF studies show that lung function decline is nonlinear and heterogeneous; using an exponential correlation structure can improve predictions of lung function decline [ 5 , 6 ]. We recently used a nonstationary Gaussian linear mixed-effects model [ 7 ] to predict rapid FEV 1 decline using data from the US Cystic Fibrosis Foundation Patient Registry (CFFPR) [ 8 ]. Specifically, we applied a nonlinear model to simultaneously fit both population- and individual-level FEV 1 decline.…”

Linear Mixed Effects Models for Non-Gaussian Continuous Repeated Measurement Data

“…Describing the time course of cystic fibrosis (CF) lung function was a compelling application, justifying the importance of introducing and relaxing assumptions of the process W i .t ij / and measurement error Z ij . By framing the prediction problem as a target function, one can use their model to form predictive probability distributions of rapid disease progression and translate probabilities for downstream informatics (Szczesniak et al, 2019). In Fig.…”

Functional data analysis and prediction tools for continuous glucose-monitoring studies

“…This model allows for estimating risk based on target functions, which we specified as glycemic excursions. Previously, this model has been used in studies of renal failure and cystic fibrosis [27]. This model captures between and within patient heterogeneities by a random intercept and stochastic process.…”

“…Methods stemming from our chosen approach have been mostly utilized for monitoring in clinical trials to compute the predictive probability of success given interim data [25]. Recently, predictive probabilities have been expanded in the context of monitoring progression toward renal failure [26] and rapid progression of lung function in cystic fibrosis [27].…”