Long-Term Trajectories
Long-term outcomes in ME/CFS are variable. Meta-analyses suggest that full recovery occurs in approximately 5% of patients, while 20–30% show some improvement, and the remainder remain stable or worsen (Brurberg et al. 2014). The model reproduces these statistics through patient-to-patient variation in parameters. Monte Carlo simulation with parameter distributions derived from population data generates a distribution of trajectories:
- Recovery (\(~\) 5%): Patients whose parameters place them near the separatrix boundary, where spontaneous fluctuations or modest interventions can push the system back to the healthy attractor
- Improvement (\(~\) 25%): Patients where damage accumulation is slow (\(k_\text{damage}\) low) and repair is sufficient to gradually reduce \(D_\text{total}\), especially with effective pacing
- Stable (\(~\) 50%): Patients at approximate damage–repair equilibrium, whose severity depends on management quality
- Progressive (\(~\) 20%): Patients where damage outpaces repair, often due to severe energy deficit limiting repair capacity or repeated perturbations (infections, overexertion) preventing equilibrium
Long-term trajectory predictions are highly sensitive to parameter values and perturbation history, both of which are poorly characterized for individual patients. The trajectory distribution above should be interpreted as a population-level statistical prediction, not as a prognosis for any individual patient. Personalized trajectory prediction requires longitudinal data collection over months to years with sufficient biomarker resolution to constrain patient-specific parameters—a capability not yet available in clinical practice.