A major goal of current pain research is to develop scientifically-based guidelines to optimize selection of analgesic drug doses and control pain. Numerous clinical studies have shown that the drug dose required to effectively alleviate pain is often highly variable, both between patients and between pain episodes in individual patients. This high variability makes it difficult to predict appropriate dosing regimens that will adequately control pain in the individual patient, The result is a ''trial and error'' approach to analgesic therapy. Experimental tools that measure analgesic response and drug disposition after administration of analgesics can be used to better predict the therapeutic effects of analgesics in individual patients. The approach defines factors that contribute to variability in therapeutic response, Pharmacokinetic-pharmacodynamic (PK/PD) modeling involves analysis of drug disposition data and drug efficacy data, Modeling depends on sensitive and reliable methods to quantify both pain and plasma (or other body fluid) levels of analgesic agents. Two major approaches have been used to quantify pain. The first utilizes subjective reports from patients, while the second employs physiological correlates of pain, such as evoked potentials. Results from PK/PD analysis that successfully identify a relationship between drug dose, drug concentration, and effect can be used to predict the effects of drug dose on analgesic effect in individuals. The ultimate goal is to provide patients with better pain relief by understanding variables that affect the analgesic concentration/effect relationship, This review examines the available pharmacokinetic-pharmacodynamic (PK/PD) data for selected opioid and nonopioid analgesics. Even though most analgesics are used clinically in multiple doses, the majority of PK/PD studies conducted to date evaluated single dose effects. Further studies with multiple doses are required to evaluate the validity of PK/PD relationships defined from single dose studies.