Disease statistical analysis

Microarrays are commercially available that contain a million SPNs that can be searched for specific gene sequences. Affymetrix (Santa Clara, CA) has developed a gene chip that can examine automatically a million SNPs at one time. SNPs do not cause disease, but they can identify segments of mutated genes that increase the risk of developing a disease. Statistical analysis is used to detect the mutated genes. 

The result of the Phase II trial is information needed to determine the effective dose and the dosing regimen of frequency and duration. Specihc chnical endpoints or markers are used to assess interaction of drug and disease. There are two types of markers definitive and surrogate. For example, in the case of cancer or hypertension, the definitive markers are mortality and stroke, respectively, and the surrogate markers may be tumor size, or cancer-associated proteins p53, TGF-a in the case of cancer, and blood pressure or cholesterol level in hypertension. Statistical analysis is carried out to evaluate the... 

In an application of this model for assessing factors that cause large interindividual variations in the clearance of theophylline (37), the authors clearly recognized its limitations and this requirement that the conclusions be tested by a prospective controlled experiment "The factors identified as important in theophylline body clearances are associations found by retrospective statistical analysis which need not imply a cause-and-effect relationship, especially where a pathophysiological or drug interaction rationale does not exist. Often these factors need further confirmation by prospective examination of cohorts of subjects with the disease or history in question."... 

Epidemiology analyzes statistical data to deiennine tlie lelationslup between a chemical and tlie exposed populatioa However, positive statistical analysis does not always translate into a cause and effect relationsltip between the chemical and disease in humans. Wlien a cause and effect relationsliip between the chemical and disease has been establislied, hirtlier statistical analysis helps define tlie upper limit of hmnan toxic risk. 

When the patients who received a drug candidate have the disease manifestations completely eradicated and experience no other effects while patients treated with a placebo have a continuation of the disease process, the evaluation is not difficult. However, that is rarely, if ever, the case, and evaluation requires detailed statistical analysis of the collected data. An ICH guide-line covers statistical issues related to the scope of clinical trials, design techniques to minimize bias, types of clinical trial designs, conduct considerations, data analysis for efficacy, evaluation of safety and tolerance, and reporting. 

Evolutionary designs were devised by Dixon and Armitage. Although the statistical analysis is rather different, they have the same objective, which is to detect a treatment effect at the earliest moment possible, using the fewest possible patients, while retaining statistical robustness. Both types are suited for exploratory clinical research and diseases which are rare. 

Cohort studies are considerably more expensive than case-control studies, and may require long periods of time before enough cases of disease or death show up in the cohorts to allow meaningful statistical analysis. 

Expression profifing coupled with appropriate statistical analysis holds promise not only in cancer classification, but also by extension to many other areas of disease research and management. Transcription profile databases maybe assembled from samples that differ by tissue source, disease state or progression, morbidity/mortahty, response to drugs and other treatments, and countless other variables. New patterns may be revealed and disease classes refined or discovered. Patients may be more finely stratified in clinical trials so that the success of treatments can be better judged and the expectation is that the diagnosis and treatment of disease will improve substantially as a result. 

Mintz GS, Popma JJ, Pichard AD et al (1995) Patterns of calcification in coronary artery disease. A statistical analysis of intravascular ultrasound and coronary angiography in 1155 lesions. Circulation 91 1959-1965... 

---