Performance evaluation in Bayesian adaptive randomization

Bayesian adaptive randomization (BAR) is an attractive approach to allocate more patients to the putatively superior arm based on the interim data while maintains good statistical properties attributed to randomization. Under this approach, patients are adaptively assigned to a treatment group based on the probability that the treatment is better. The basic randomization scheme can be modified by introducing a tuning parameter, replacing the posterior estimated response probability, setting a boundary to randomization probabilities. Under randomization settings comprised of the above modifications, operating characteristics, including type I error, power, sample size, imbalance of sample size, interim success rate, and overall success rate, were evaluated through simulation. All randomization settings have low and comparable type I errors. Increasing tuning parameter decreases power, but increases imbalance of sample size and interim success rate. Compared with settings using the posterior probability, settings using the estimated response rates have higher power and overall success rate, but less imbalance of sample size and lower interim success rate. Bounded settings have higher power but less imbalance of sample size than unbounded settings. All settings have better performance in the Bayesian design than in the frequentist design. This simulation study provided practical guidance on the choice of how to implement the adaptive design. ^

Assessment of the performance of 3m 3500 organic vapor monitors over extended sampling durations

The purpose of this study was to assess the accuracy and precision of airborne volatile organic compound (VOC) concentrations measured using passive air samplers (3M 3500 organic vapor monitors) over extended sampling durations (9 and 15 days). A total of forty-five organic vapor monitor samples were collected at a State of Texas air monitoring site during two different sampling periods (July/August and November 2008). The results of this study indicate that for most of the tested compounds, there was no significant difference between long-term (9 or 15 days) sample concentrations and the means of parallel consecutive short-term (3 days) sample concentrations. Biases of 9 or 15-day measurements vs. consecutive 3-day measurements showed considerable variability. Those compounds that had percent bias values of <10% are suggested as acceptable for long-term sampling (9 and 15 days). Of the twenty-one compounds examined, 10 compounds are classified as acceptable for long-term sampling; these include m,p-xylene, 1,2,4-trimethylbenzene, n-hexane, ethylbenzene, benzene, toluene, o-xylene, d-limonene, dimethylpentane and methyl tertbutyl ether. The ratio of sampling procedure variability relative to variability within days was approximately 1.89 for both sampling periods for the 3-day vs. 9-day comparisons and approximately 2.19 for both sampling periods for the 3-day vs. 15-day comparisons. Considerably higher concentrations of most VOCs were measured during the November sampling period compared to the July/August period. These differences may be a result of varying meteorological conditions during these two time periods, e.g., the differences in wind direction, and wind speed. Further studies are suggested to further evaluate the accuracy and precision of 3M 3500 organic vapor monitors over extended sampling durations. ^

The Need for Balanced Methods of Measuring Performance

Michelle Rhee weighs in on measuring success in pubic education.