Search

In this paper we describe a decision analytic approach to doing a futility analysis of a phase II study on a drug under development. For this analysis we build a model that estimates the expected monetary value of the drug and the expected costs of development. These statistics are used to help decide whether to continue the study.

In this paper we present a decision analytic approach to determining sample size in a set of Phase III drug efficacy trials sponsored by a pharmaceutical company. In this approach we built a model to predict the expected net present value of the drug at varying sample sizes. We then chose the sample size that maximizes that value. We took into consideration effects that sample size had on the probability of approval, the cost of the studies, the time it takes the drug to reach the market, and a variety of other factors. We found that increasing the sample size increases the chance that the drug will be approved. On the other hand increasing the sample size increases the cost of the studies and the time it will take for the drug to reach the market. The model weighs these factors to produce the expected net present value.

In this paper we describe a decision analysis that was designed to help a pharmaceutical company decide whether to continue development of a drug after the completion of a Phase II program. We use a Bayesian hierarchical model to integrate data from three Phase II studies, data from related studies reported in the literature, and prior assumptions about the pattern of response across dose and dosing frequency. The product of this model is a posterior distribution of the efficacy and tolerance parameters. This distribution is then combined with marketing data in a decision model. The decision model includes a second (post-Phase III) decision point. The main goal of this report is to present the methodology of this decision analysis. There is also some limited presentation of results.

Background: Trafermin (basic fibroblast growth factor) has been shown to reduce infarct volume in acute ischemic stroke models, and to promote functional recovery and new synapse formation when given to animals with completed cerebral infarction. A previous study in acute stroke patients suggested that trafermin was safe and well tolerated when given over a 3-hour period over a wide dose range. Methods and Results: Double-blind, parallel group, placebo-controlled trial of a single 24-hour intravenous infusion of trafermin. Patients having onset of stroke symptoms within 6 h and a baseline score of ≧7 on the NIH Stroke Scale (≧2 motor) were randomized to receive 5 or 10 mg of trafermin or placebo intravenously infused over 24 h. The primary efficacy outcome was a categorized combination of the Barthel and Rankin scales assessed at 90 days. A total of 286 patients had been enrolled at 55 sites in 11 countries when the sponsor directed that enrollment be stopped because an interim analysis of efficacy data predicted too small a chance of demonstrating a statistically significant benefit after recruitment of the planned 900 patients. The 5-mg group showed a slight but nonsignificant advantage over placebo (OR 1.2, 95% CI 0.72–2.00, p = 0.48); the 10-mg group showed a nonsignificant disadvantage (OR 0.74, 95% CI 0.44–1.22, p = 0.24). Mortality rates at 90 days were 17% in the 5-mg group, 24% in the 10-mg group and 18% in the placebo group. Treatment with trafermin was associated with an increased leukocytosis and a decrease in blood pressure: mean decrease in systolic blood pressure from baseline was 19 mm Hg in the 5-mg group, 22 mm Hg in the 10-mg group and 8 mm Hg in the placebo group. In a post hoc subgroup analysis, patients in the 5-mg group treated more than 5 h after the onset of symptoms showed an apparent advantage over placebo (OR 2.1, 95% CI 1.00–4.41, p = 0.044; after age adjustment: OR 1.9, 95% CI 0.91–4.13, p = 0.08). Conclusions: With the proper treatment regimen, trafermin can likely be given safely to stroke patients. The 5-mg dose showed a trend toward a treatment advantage. The ideal time window for this agent may exceed 5 h. This may open new avenues for acute stroke therapy, aiming at enhancing recovery mechanisms rather than immediate neuroprotection.

In recent years, many researchers have been interested in performing futility analyses at interim points during clinical trials in order to terminate futile trials. To implement such an analysis, the researcher must decide what percentage of the total number of patients must have completed the trial before the interim analysis is done. This percentage is usually between 20% and 80%. We examined the relationship between the percentage chosen and the probability of stopping the trial, given a parameter value associated with the treatment effect. The results of this examination will help the researcher plan when to do a futility analysis.

Compounds active at the serotonin (5-HT)1A receptor (mostly azapirones) have shown some evidence of antidepressant effect. We report here the results of an antidepressant trial with zalospirone, a novel cyclic imide with 5-HT1A partial agonist activity. Two hundred eighty-seven outpatients (mean age 44 years, 55% men, 45% nonfertile women) who met criteria for unipolar major depression with a minimum 21-item Hamilton Rating Scale for Depression (HAM-D) score of 20 were randomly assigned to receive 6 weeks of double-blind treatment with either placebo or one of three fixed doses of zalospirone (6, 15, or 45 mg/day), administered three times daily. The high dose (45 mg) of zalospirone produced a significant antidepressant effect compared with placebo from week 2 on with mean improvement (change from baseline) in HAM-D total score of 12.8 versus 8.4 (p < 0.05) at week 6. Clinical improvement with the high dose of zalospirone was consistent across all outcome measures, however, only in the observed cases and not the last-observation-carried-forward analyses. Improvement with the 6-mg or 15-mg doses was greater than that with placebo, but not significantly so, suggesting a dose-response effect. Although the 45-mg dose of zalospirone seemed to have significant antidepressant efficacy, it was not well tolerated. Dizziness and nausea were noted in almost half of the patients, and by week six, 51% of patients in the high-dose group had dropped out. Whether or not tolerance to this high dose might be improved by gradual drug titration, only future research can answer.

This paper reviews the educational and emotional impact of restrictive (isolated) educational settings on children who are in the lowest quartile of achievement. The basic arguments for and against separation of slower students (both those classified as handicapped and those not so classified) are reviewed and some of the more prominent legal cases discussed. Various educational settings are defined and a position is developed as to how setting operates. The main body of the paper examines the results of studies in both special and compensatory education. As a result of the extensive review, it is asserted that the variables which are important for successful student outcomes can occur in most settings, and that for ethical reasons the least restrictive environment is preferred. In summary, it is the issue of effective practices, not the setting, that deserves the attention of educators.

Data on important methodological problems in assessing support networks is presented. Stability (over 1 year) and construct validity were examined in two groups of mothers of preschool children. One group of mothers had children with a physical handicap (n = 25); the other group had normally developing children (n = 44). When their children were 4 years of age and again at 5 years of age, mothers' support networks were assessed with the Pattison Psychosocial Kinship Inventory (a semistructured interview) and a log diary. Results indicated considerable stability in mothers' support networks over the 1-year period. In addition, agreement across methods was fairly high despite marked differences in format, providing evidence for the construct of support networks. Further evidence for construct validity was provided by generally low correlations between support network variables and other measures of social ecology. Findings were roughly similar for both groups of mothers indicating general replicability of findings.