Clinical Decision Support
Translation of models into clinical practice requires decision support tools that present model outputs in clinically actionable formats.
1 Dashboard Concept
A model-based clinical dashboard would display:
- Current state assessment: estimated positions along the energy, immune, neuroendocrine, and autonomic axes, derived from the most recent biomarker panel
- Trajectory projection: predicted 3–6 month trajectory under current management, with uncertainty bounds
- Treatment comparison: side-by-side predicted outcomes for candidate interventions
- Alert indicators: flags for approaching thresholds (e.g., damage rate accelerating, energy budget shrinking, immune markers trending upward)
Such a dashboard requires automated model fitting, fast simulation, and validated uncertainty quantification—capabilities that exist in related fields (pharmacometrics, systems biology) but have not yet been adapted for ME/CFS.
2 Treatment Algorithms
Model-derived treatment algorithms formalize clinical reasoning as decision trees parameterized by biomarker values. A simplified example:
- If \(\alpha_\text{CI} < 0.6\) (severe metabolic impairment): prioritize mitochondrial support (CoQ10, NAD⁺ precursors, d-ribose)
- If \(\mathbf{C}_\text{pro} > 2 \times\) healthy reference: prioritize immunomodulation (LDN, anti-inflammatory agents)
- If \(\text{HR}_\text{standing} - \text{HR}_\text{supine} > 30\) bpm: prioritize autonomic support (volume expansion, compression, pyridostigmine (Raj et al. 2005))
- If multiple criteria met: combination therapy targeting dominant pathway first, adding sequential interventions with 4–6 week monitoring intervals
These algorithms are model-informed but require clinical judgment for implementation. They do not replace physician decision-making but provide a structured framework for the complex multi-system reasoning that ME/CFS management demands.