Pooling different studies

Reconfiguration of Data. Drug safety data from different sources are often pooled or combined in databases. Reasons for combining data vary. In the case of premarketing studies, data from different sites are routinely combined because one site may not be able to recruit enough patients for a study. Data from different studies are often combined to increase sample size and therefore statistical power for detecting an uncommon adverse event. 

Carbon represents about 50% of the total oven-dried biomass present in forests [32]. Estimation of carbon pools in forests necessarily involves studying the different strata of biomass present in them. In the different studies, the following carbon pools and variables were measured ... 

Data are pooled from different studies and are not necessarily comparable. 

The primary results of economic evaluations usually is a comparison of average, or pooled, differences in costs and differences in effects among patients who received the therapies under study. It is an open question, however, whether pooled results are representative of the results that would be observed in the individual centers or countries that participated in the study. In some, the therapy may provide good value for the costs, whereas in others it may provide poor value. Three reasons commonly cited for these differences are differences in practice patterns (i.e., medical service use), differences in absolute and relative prices for medical service use (i.e., unit costs), and differences in underlying morbidity/mortality patterns in different centers and countries. 

Accuracy is a central theme in any analytical process and its pursuit can no longer be the work of a single analytical chemist but requires comparative studies involving many methodologies and many experts. It is sensible to expect that pooling different kinds of data (methods and users) will lead to better estimates of the truth, even if some of the contributions are more reliable than others. 

The PPK approach can allow one to combine heterogeneous types of data from varying sources. For example, one could pool data from several different studies, study centers, variable biomatrices (plasma plus serum), intensely plus sparsely sampled data, or experimental plus observational data. The combining of differing data sets often increases the power to identify multicompartment or nonhnear models, to incorporate additional covariates, or to gain precision in the estimation of the model. 

Since the concentration of SM-C/IGF-I has traditionally been expressed in U/ml, with I U/ml arbitrarily defined as the activity in a pool of normal adult plasma, it has been difficult to compare in absolute terms the values obtained in different laboratories. As purified peptides have become available, it is now possible to express both SM-C/IGF-I and IGF-II in mass units, generally ng/ml. In a number of laboratories which report in mass units or provide conversion foctors between U/ml and ng/ml, the level of SM-C/IGF-I in apparently healthy adults is generally about 200-300 ng/ml, although some laboratories have reported different mean values in different studies. Some typical mean values and ranges (either presented as published or calculated from data provided) are shown in Table 2. In recent years reference ranges for IGF-II have also become established. As seen in Table 2, mean values are 2 to 3 times higher than for SM-C/IGF-I. 

The interpretation of studies that have pooled different treatments... 

From an analysis standpoint, often only the trials with all common doses are included, which can exclude important information. One may also crudely pool patients at the same dose level across studies, but the drawback is that it breaks down the randomization. Bayesian methodology has been shown to be useful in the context of indirect and mixed treatment comparison methods, to combine information from different therapies in different studies in order to make treatment effect inferences, but instead of modeling differenf dose arms in different studies, we extend the methodology to allow for assessment of the dose-response relationship across multiple clinical trials. 

This section presents the two most widely used statistical methods for meta-analysis, namely, the fixed effecfs model and the random effects model. In addition, we want to emphasize that an analysis based on crude pooling of adverse event numbers across different studies to compare treatment groups should be avoided, as the analysis is vulnerable to the mischief of Simpson s paradox (Chuang-Stein and Beltangady, 2011). 

The results from different studies carried out independently, or in collaboration, can be combined statistically by meta-analysis. Meta-analysis is the quantitative cumulation and analysis of descriptive statistics across studies (Hunter et al, 1982). Although methods for combining probabilities across studies have been available for some years, the methods and techniques of meta-analysis are recent, and have been developed within the last decade, mostly for experimental or quasi-experimental studies. The basic statistic of meta-analysis is d, which is computed as the difference between group means, divided by the pooled within-group standard deviation. 

Sta.bilizers. Cyanuric acid is used to stabilize available chlorine derived from chlorine gas, hypochlorites or chloroisocyanurates against decomposition by sunlight. Cyanuric acid and its chlorinated derivatives form a complex ionic and hydrolytic equilibrium system consisting of ten isocyanurate species. The 12 isocyanurate equilibrium constants have been determined by potentiometric and spectrophotometric techniques (30). Other measurements of two of the equilibrium constants important in swimming-pool water report significantly different and/or less precise results than the above study (41—43). A critical review of these measurements is given in Reference 44. 

---