Bayes without priors

Subjectivity of pre-test probability value: controversies over the use of Bayes' Theorem in medical diagnosis

This article discusses the use of Bayes' Theorem in medical diagnosis with a view to examining the epistemological problems of interpreting the concept of pre-test probability value. It is generally maintained that pre-test probability values are determined subjectively. Accordingly, this paper investigates three main philosophical interpretations of probability (the "classic" one, based on the principle of non-sufficient reason, the frequentist one, and the personalistic one). This study argues that using Bayes' Theorem in medical diagnosis does not require accepting the radical personalistic interpretation. It will be shown that what distinguishes radical and moderate personalist interpretations is the criterion of conditional inter-subjectivity which applies only to the moderate account of personalist interpretation.

Bayesian methodology for the design and interpretation of clinical trials in critical care medicine: a primer for clinicians

OBJECTIVES

To review Bayesian methodology and its utility to clinical decision making and research in the critical care field.

DATA SOURCE AND STUDY SELECTION

Clinical, epidemiological, and biostatistical studies on Bayesian methods in PubMed and Embase from their inception to December 2013.

DATA SYNTHESIS

Bayesian methods have been extensively used by a wide range of scientific fields, including astronomy, engineering, chemistry, genetics, physics, geology, paleontology, climatology, cryptography, linguistics, ecology, and computational sciences. The application of medical knowledge in clinical research is analogous to the application of medical knowledge in clinical practice. Bedside physicians have to make most diagnostic and treatment decisions on critically ill patients every day without clear-cut evidence-based medicine (more subjective than objective evidence). Similarly, clinical researchers have to make most decisions about trial design with limited available data. Bayesian methodology allows both subjective and objective aspects of knowledge to be formally measured and transparently incorporated into the design, execution, and interpretation of clinical trials. In addition, various degrees of knowledge and several hypotheses can be tested at the same time in a single clinical trial without the risk of multiplicity. Notably, the Bayesian technology is naturally suited for the interpretation of clinical trial findings for the individualized care of critically ill patients and for the optimization of public health policies.

CONCLUSIONS

We propose that the application of the versatile Bayesian methodology in conjunction with the conventional statistical methods is not only ripe for actual use in critical care clinical research but it is also a necessary step to maximize the performance of clinical trials and its translation to the practice of critical care medicine.

Trichloroethylene cancer epidemiology: a consideration of select issues

A large body of epidemiologic evidence exists for exploring causal associations between cancer and trichloroethylene (TCE) exposure. The U.S. Environmental Protection Agency 2001 draft TCE health risk assessment concluded that epidemiologic studies, on the whole, support associations between TCE exposure and excess risk of kidney cancer, liver cancer, and lymphomas, and, to a lesser extent, cervical cancer and prostate cancer. As part of a mini-monograph on key issues in the health risk assessment of TCE, this article reviews recently published scientific literature examining cancer and TCE exposure and identifies four issues that are key to interpreting the larger body of epidemiologic evidence: a) relative sensitivity of cancer incidence and mortality data ; b) different classifications of lymphomas, including non-Hodgkin lymphoma ; c) differences in data and methods for assigning TCE exposure status ; and d) different methods employed for causal inferences, including statistical or meta-analysis approaches. The recent epidemiologic studies substantially expand the epidemiologic database, with seven new studies available on kidney cancer and somewhat fewer studies available that examine possible associations at other sites. Overall, recently published studies appear to provide further support for the kidney, liver, and lymphatic systems as targets of TCE toxicity, suggesting, as do previous studies, modestly elevated (typically 1.5-2.0) site-specific relative risks, given exposure conditions in these studies. However, a number of challenging issues need to be considered before drawing causal conclusions about TCE exposure and cancer from these data.

Empirical-Bayes adjustment improved conventional estimates in postmarketing drug-safety studies

OBJECTIVES

Empirical-Bayes adjustment was used to investigate the effects of antibacterials on risk of arrhythmia in a large cohort study based upon automated record linkage.

STUDY DESIGN AND SETTING

The Prescription Drug Database of the Varese Province, Italy, was used to form a population-based cohort of patients who received at least one prescription of antibacterials for systemic use. All cohort members were followed-up from the beginning of each antibacterial therapeutic cycle up to 60 days afterward, identifying the first prescription of an antiarrhythmic drug and/or the first admission for arrhythmia. Standardized incidence ratios for eight classes of antibacterials, as well as for the specific antibacterial drugs, were estimated using both conventional methods and empirical-Bayes (EB) adjustment.

RESULTS

Using EB adjustment, overall precision was greatly enhanced compared to the conventional analysis. Statistical evidence of drug-induced dysrhythmic events was found for 17 drugs mostly belonging to Macrolides and Quinolones. Strongest proarrhythmic effects were found for Ciprofloxacin (standardized incidence ratio [SIR]: 1.98; 95% CI: 1.67, 2.34) and Erythromycin (SIR: 1.91; 95% CI: 1.20, 2.88).

CONCLUSION

Empirical-Bayes may mitigate some problems inherent the conventional analysis. Macrolides and Quinolones were associated with the highest number of positive signals of dysrhythmic events. This requires further investigation by other types of epidemiologic studies.