Selected Topics in Statistical Analysis of Clinical Research

Abstract

Statistical analysis usually is employed in the evaluation of clinical research studies. This paper reviews and makes recom mendations in three areas frequently overlooked in the conduct of clinical research: power analysis, specification of a priori research hypotheses, and Bayesian analysis. Power analysis determines the number of subjects required to conduct a mean ingful study and should be performed during the planning phase. Research and null hypotheses are essential elements of research design and should be specified prior to statistical analysis. Bayesian analysis can be used both to evaluate diag nostic tests and as an alternative to traditional statistical approaches for testing multiple hypotheses. Application of these methods is described and clinical examples are provided. The discussion is nontechnical and is directed toward the clinical researcher. (J Geriatr Psychiatry Neurol 1989;2:34-40).

Get full access to this article

View all access options for this article.

References

Hahn GJ : More intelligent statistical software and statistical expert systems: Future directions. Am Stat 1985;39:1-7.

Bailar JC , Mosteller F. : Guidelines for statistical reporting in articles for medical journals: Amplifications and explanations. Ann Intern Med 1988;108:266-273.

Committee on Professional Misconduct : Report of Guidelines for Investigators in Scientific Research. Boston, MA, Harvard Medical School, 1988.

Andrews FM , Klem L. , Davidson TN , et al: A Guide for Selecting Statistical Techniques for Analyzing Social Science Data, ed 2. Ann Arbor, MI, University of Michigan, 1981 .

Ahern DK , Baer L. : StatChoice: Development of a program for statistical consultation. Behav Ther 1985;8:190.

Browner WS , Newman TB : Are all significant p values created equal? The analogy between diagnostic tests and clinical research. JAMA 1987;257:2459-2463.

Gigerenzer G. , cited by Cohen J. : Statistical Power Analysis for the Behavioral Sciences, ed 2. Hillsdale, NJ , Lawrence Erlbaum Assoc, 1988.

Scott JA , Phillips WC , Blasczcynski GM : Statistics for diagnostic procedures. AJR 1983;141:409-411.

McClarty JW : How many subjects are required for a study. IRB: A Review of Human Subjects Research 1987;9:1-3.

10.

Cohen J. : The statistical power of abnormal-social psychological research: A review . Abnorm Soc Psychol, 1962;65:145-153.

11.

Anderson RB : Stat Power. An Apple Computer Program. Cam-bridge, MA, Abt Associates, 1981.

12.

Baer L. , Ahern DK : StatChoice: The Statistical Consulting Program. Littleton, MA, PSG Publishing Co, 1988.

13.

Borenstein M. , Cohen J. : Statistical Power Analysis: A Computer Program . Hillside, NJ, Lawrence Erlbaum Assoc, 1988.

14.

Dallal GE : Design. Evanston, IL, Systat, 1987.

15.

Winer BJ : Statistical Principles in Experimental Design, ed 2. New York, McGraw Hill, 1971.

16.

McNeil BJ , Keeler E. , Adelstein SJ : Primer on certain elements of medical decision making. N Engl J Med 1975;293:212-215.

17.

Phillips WC , Scott JA , Blasczcynski G. : How sensitive is "sensitivity"; How specific is "specificity"? AJR 1983;140:1265-1270.

18.

Diagnostic and Statistical Manual of Mental Disorders, ed 3, revised. Washington, DC, American Psychiatric Association, 1985.