A therapy-driven approach to designing personalized pain care in sickle cell disease

Abstract Background: Chronic pain in Sickle Cell Disease (SCD) is not one-size-fits-all. It stems from a mix of biological, psychological, and social factors, yet treatment strategies remain largely reactive and generalized. Traditional efforts to personalize care have focused on sorting patients into pain subtypes based on complex data and then matching therapies afterward. While informative, this process can be difficult to apply at the bedside. We propose a different approach: starting with the therapies we believe can help and working backward to define the patient profiles most likely to respond. This therapy-first framework centers feasibility, clinical impact, and real-world translation. Methods: We outline a prospective framework organized around three therapy bundles commonly used in SCD pain care: (1) inflammation-focused treatments (e.g., anti-inflammatories, transfusions), (2) neurocentral strategies (e.g., neuromodulators, CBT informed by neuroimaging), and (3) psychosocial support (e.g., trauma-informed therapy, music or nutrition interventions). For each bundle, we identify the relevant features to collect—such as quantitative sensory testing (QST), actigraphy, ecological momentary assessment (EMA), inflammatory biomarkers, ASL and resting-state fMRI, cognitive function, and whole-person assessments like mood, trauma history, and fatigue. Rather than clustering patients without context, we use machine learning to explore which features best predict a positive response to each therapy bundle—effectively shaping phenotypes around real treatment options. Results: This proposed study aims to show that it's possible to build meaningful, therapy-matched patient profiles using this reversed design. Instead of defining subtypes and hoping for treatment matches, we define what works and find out who it's most likely to help. We anticipate identifying distinct and stable patient phenotypes that align with known mechanisms and offer clear next steps for care. Validity testing will focus on reproducibility, predictive value, and alignment with existing clinical observations. Conclusion: Personalizing pain care in SCD requires more than just better data—it needs better alignment between mechanisms and treatments. By starting with the therapies and building backwards, this framework creates a more practical path to precision care. It prioritizes real-world use while still honoring the complexity of SCD pain biology. This approach may also serve as a model for other chronic pain conditions where one-size-fits-all medicine continues to fall short.