Responses to pharmacotherapy for acute and chronic pain are highly variable, and efficacy is often compromised by some form of toxicity. To increase our understanding of complexities of pharmacotherapy, the authors discuss an approach to identify analgesic responder subgroups and predictors of response. Additionally, analgesic efficacy and toxicity were combined in a single risk-benefit index (utility function) to quantify the probability of side effects in high- vs low-analgesic responders. The subgroup analysis consists of a mathematical description or time series analysis, mixture analysis, and covariate analysis. Applied to ketamine treatment of chronic pain in complex regional pain syndrome and capsaicin 8% patch treatment in post-herpetic neuralgia patients, the analyses yielded homogenous subgroups that differed in distribution frequency between treatments. For capsaicine, a high variability in pretreatment pain reporting was associated with a high probability of falling into a full analgesic subgroup, irrespective of treatment. The utility function was applied to opioid-induced analgesia and respiratory depression. An important observation was that, irrespective of dose, low-analgesic responders to fentanyl had a greater probability of respiratory depression than analgesia while the reverse was true for high-analgesic responders. These data show dissociation between 2 μ-opioid end-points and explain the danger of treating poor analgesic responders with increasingly higher opioid doses. Apart from being valuable in drug development programs, the outlined approach can be used to determine the choice of drug and dose in the treatment of pain in patients with potent and toxic analgesics.