Background and Objectives: Despite management with opioids and other pain modifying therapies, neuropathic pain continues to reduce the quality of life and daily functioning in HIV infected individuals. Cannabinoid receptors in the central and peripheral nervous systems have been shown to modulate pain perception. We conducted a clinical trial to assess the impact of smoked cannabis on neuropathic pain in HIV. Design and Methods: This was a phase II, double-blind, placebo-controlled, crossover trial of analgesia with smoked cannabis in HIV-associated distal, sensory predominant polyneuropathy (DSPN). Eligible subjects had neuropathic pain refractory to at least two prior analgesic classes; they continued on their pre-study analgesic regimens throughout the trial. Regulatory considerations dictated that subjects smoke under direct observation in a hospital setting. Treatments were placebo and active cannabis ranging in potency between 1 and 8 % delta-9-tetrahydrocannabinol (THC), four times daily for five consecutive days during each of two treatment weeks,

Medical laboratory data are often censored, due to limitations of the measuring technology. For pharmacokinetics measurements and dilution-based assays, for exam-ple, there is a lower quantification limit, which depends on the type of assay used. The concentration of HIV particles in the plasma is subject to both lower and upper quantification limit. Linear and nonlinear mixed effects models, which are often used in these types of medical applications, need to be able to deal with such data issues. In this paper we discuss a hybrid Monte Carlo and numerical integration EM algorithm for computing the maximum likelihood estimates for non-linear mixed models with cen-sored data. Our implementation uses an efficient block-sampling scheme, automated monitoring of convergence, and dimension reduction based on the QR decomposition. For clusters with up to two censored observations numerical integration is used instead of Monte Carlo simulation. These improvements apply to both linear and nonlinear mixed effects models with censored data and lead to a several-fold reduction in com-putation time. We illustrate the algorithm on two applications from HIV/AIDS: a treatment trial and an study of acute HIV infection; some of the data in these studies, HIV-1 RNA viral load, are left- and right-censored, respectively. The Monte Carlo EM is evaluated and compared with existing methods via a simulation study.