Integrating pharmacokinetic modeling and wearable-derived physiological data to assess pain intensity in patients with advanced urologic cancer.

BACKGROUND

While opioid analgesics are widely used to manage moderate-to-severe cancer-related pain, dosing is mostly based on subjective pain reports rather than objective physiological indicators, potentially contributing to inconsistent dosing and suboptimal pain control. This study investigated whether model-predicted oxycodone serum concentrations could guide opioid dosing and whether wearable-derived heart rate and step count, combined with these predictions, could complement conventional numerical rating scale (NRS) pain assessments.

METHODS

Sixteen patients with advanced genitourinary cancer receiving oxycodone were prospectively monitored. Clinically collected serum concentrations were used to simulate individual pharmacokinetic profiles. Pain intensity was assessed using the NRS, and physiological parameters were recorded using wearable devices. Linear mixed-effects models were constructed to evaluate the associations between simulated serum oxycodone concentrations, NRS, and physiological variables.

RESULTS

In the linear mixed-effects model, simulated serum oxycodone concentrations were significantly associated with lower NRS scores (β = -0.26, P = .0177). Although neither heart rate nor step count independently predicted NRS, incorporating heart rate into the model improved the overall fit. Thus, heart rate may capture pain-related physiological responses not fully explained by oxycodone concentration alone, thereby enhancing the explanatory power of the model.

CONCLUSIONS

Model-based simulated serum oxycodone concentrations may serve as an objective reference for individualized pain assessment and opioid titration. Combining wearable-derived physiological signals, particularly heart rate, with model-predicted oxycodone concentrations may further improve the precision and adaptability of pain management in palliative care.