MEDICAL STATISTICS

Editor's Note: Few introductory words will suffice. Prof. Dimitar Sepetliev /d. 2006/ has been a lecturer in medical statistics for more than thirty years. His textbooks were a handy tool for many a aspiring scholars - both, in Bulgaria and abroad. In the 70s and 80s of the past century, he was teaching statistics at the Moscow International Courses for Health Administrators /i.e., with support from WHO/. Simultaneously, he was a collaborator in statistical techniques and research methodologies at the High Institute of Economics "Karl Marx" in Sofia. At the brink of the democratic changes from 10 November 1989 - comprehensively, Prof. D. Sepetliev has become a Member of the National Parliament. He was instrumental in applying new electoral techniques for voting and balloting in the National Assembly, ditto.

The development of statistics has been like that of language itself. Its origins are ancient. For instance, the use of the mean or average was well known at the time of Pythagoras, and mention of statistical surveys was made in biblical times. Like language, statistics developed gradually where it was needed. As society became more complex, there developed more demand for accurate summary statements and inference made in numerical form. The discipline of statistics has in recent years gained momentum both in its mathematical development and through its many applications in new fields. The modern development of statistics began in the sixteenth century, when governments of various western European countries became interested in collecting information about their citizens. By the seventeenth century, surveys that closely resemble our modern census were already being conducted.

By that time, insurance companies were beginning to thrive and were already compiling mortality tables to determine their life insurance rates. Different vested interests were slowly turning toward enterprises that necessitated the treatment of data and that demanded the use of statistics. The mathematical basis of statistics is certainly not the subject matter of this book, but its existence, like the engine of an automobile, is vital. Thus Isaac Newton (1642—1727), whose contribution to the invention of calculus was an outstanding event in mathematics, was perhaps the most necessary figure for the development of modern statistics, though Newton could scarcely have heard of the subject. Other mathematicians whose contributions have been primarily in the field of pure mathematics have done more for the
development of statistics indirectly than many of those whose names are associated specifically with the field of statistics itself. Perhaps the two most prominent ones are Abraham DeMoivre (1667—1754) and Carl Gauss (1777—1855).

As for the statisticians themselves, Adolph Quetelet (1796—1874) a Belgian, was the first to apply modern methods to collected data. Quetelet is some time referred to as the father of modern statistics, less because of his contributions than because of his continued emphasis on the importance of using statistical methods. After studying with the best-known mathematicians of his day, Quetelet established a Central Commission for Statistics which became a model for similar organizations in other countries.

Unexpected as it sounds, Florence Nightingale (1820—1910) was an ardent proponent of the use of statistics all of her life. She argued that the administrator could be successful only if he were guided by statistical knowledge and that both the legislator and the politician often failed because their statistical knowledge was insufficient.

Two other important contributors to statistics were the Englishmen Sir Francis Galton (1822—1911) and Karl Pearson (1857—1936). Galton, a cousin of Charles Darwin, became deeply interested in the problem of heredity, to which he soon applied statistical tools. Among other things, he developed the use of percentiles. Pearson made many statistical discoveries, and both Galton and Pearson contributed greatly to the development of correlation theory, which we shall consider in detail later.

The most prominent contributor to the field of statistics in the twentieth century has been Sir Ronald Fisher (1890—1962). Fisher made continuous contributions from 1912 to 1962 and many of them have had great impact on contemporary statistical procedures.

The twentieth century has seen the birth and growth of formal statistics instruction in the United States. At the turn of the century only a handful of statistics courses were given in all the colleges in America. In fact, most of these were given in economics departments and by instructors whose focus made them unique among their colleagues. During the first thirty years of this century the emphasis on applications of statistics to problems of psychology slowly increased; but it should be noted that during this period psychology itself had a shallow status and was frequently considered only as a branch of philosophy.

Shortly before World War II the number of applications of statistical methods in the social sciences began to increase. The number of surveys of all kinds increased, and the need to interpret data in psychology and education made it necessary for workers to have at least a basic understanding of statistics. One by one, statistics courses were added as requirements in psychology departments and in schools of education. Today the worker in any of the human behavior fields, such as psychology, sociology, or anthropology, is expected at least to have what is called statistical literacy. In fact, it is virtually impossible to major in any of these fields without having to take at least one statistics course.

One reason for the rapid growth of statistics in recent years is the increasing ease of processing large masses of data. Modern electronic computers make it possible to analyze in a short time huge collections of data that would have been entirely unmanageable just a few years ago. Such data can now be analyzed and the results made available for public and professional consumption while they are still pertinent to the current situation. If you should ever have to make an extensive statistical survey, you will undoubtedly enlist the aid of modern methods of computation. Familiarity with a desk calculator is useful for less extensive computations.

One implication of the existence of rapid computational devices is that there is much less need for an emphasis on this aspect in the teaching and learning of elementary statistical concepts. The time is much more profitably spent in furthering a real understanding of the ideas involved. On the other hand, a certain amount of practice in working with the various statistical measures is actually helpful in developing this understanding. Problems in the form of case studies are designed for this purpose.

(i). This is not yet a computer and such forms for disease registration have been used for at least 500 /five hundred/ years. Also, vital statistics made use of it /i.e., births, deaths, marriages, divorces, etc/. The mechanism of "coding" is achieved through perforating a card on its margins by a stapler. Then a lever is broached through a patch of cards and shaken. Those cards with a variable under investigation fall out from the patch. More direct and convenient methods for reading statistical information have evolved with the advent of computers and those are now generally preferred instead of the punched cards teqnique.

Addendum: We could find, recently, a paper published by D. Sepetliev in the series "Public Health in Europe - № 1: Health Planning and Organization of Medical Care"