Why Model ME/CFS?
The complexity of ME/CFS pathophysiology motivates formal modeling for four reasons, each addressing a distinct limitation of verbal or qualitative reasoning.
1 Managing Multi-System Complexity
Verbal descriptions of ME/CFS pathophysiology—such as “immune activation drives neuroinflammation, which disrupts autonomic function, which impairs perfusion, which worsens energy deficits”—are inherently ambiguous about magnitudes, timescales, and thresholds. A mathematical model forces explicit specification of these quantities. When Naviaux and colleagues identified a hypometabolic state in ME/CFS patients through metabolomics (Naviaux et al. 2016), the question was not merely whether metabolic pathways are disrupted, but how much disruption is required to produce the observed clinical phenotype and whether that disruption is self-sustaining. Only quantitative models can answer such questions rigorously.
2 Generating Testable Predictions
A well-specified model generates predictions that extend beyond the data used to construct it. For instance, the IDO metabolic trap hypothesis proposed by Phair (Phair, Davis, and Kashi 2019) was formulated as a mathematical model of bistability in tryptophan metabolism. The model predicts specific dose–response relationships for tryptophan supplementation and identifies parameter regimes under which the metabolic trap is reversible. These predictions are testable independently of the original hypothesis, providing a rigorous path toward confirmation or refutation.
3 Optimizing Interventions
Treatment of ME/CFS currently relies on empirical trial-and-error. Models that capture the dynamics of disease processes can, in principle, identify optimal intervention points—the system components whose modification yields the greatest clinical improvement for the least perturbation. The energy envelope theory (L. A. Jason et al. 2012) (L. Jason et al. 2009) is an informal example: it identifies energy expenditure as the critical variable to control and proposes pacing as the intervention. A formal model can extend this reasoning to pharmacological interventions, predicting which combinations of treatments target complementary pathways.
4 Personalized Medicine
ME/CFS is heterogeneous. Patients differ in symptom profiles, severity trajectories, biomarker patterns, and treatment responses (Hornig et al. 2015) (Montoya et al. 2017). A parameterized model can, in principle, be fitted to individual patient data—metabolomic profiles, immune markers, autonomic function tests—to generate patient-specific predictions. This is the logic underlying precision medicine approaches, though current data availability limits practical implementation (see Data Requirements).