Ongoing public health stuff.

Having once more deftly managed my usual pathetic detractors, and gently further bonded with the man crush fellows, it's back again to shameless showing off, under the pretext of unsolicited public education. We've covered, in only a few Blogs, most of the ideas underlying basic, and even more advanced, Epid. It's origins, purposes, development, ground level technical features, and much more, were covered, admittedly only in a cursory fashion. But careful readers will have gotten the main points.
Of course, in these days of the Chinese Wuhan wet markets virus, many will want to know what Epid reveals about this latest gift to the world from the one party Han state. In doing so, many of the basic concepts which were covered earlier in the lecture series enter in.
Now we cover various features of statistical analysis, showing how closely related are medicine, Epid, public health and biostatistics.This actually is an extension of calculating the results of laboratory experimentation. It's rare to compare just two mice, or viral tissue culture Petri plates, for the differences after different exposures to experimental variables. And ideally, many attempts are made to have these, before the experiment, to be identical. Strains of laboratory animals are bred, at great expense, to be as nearly identical as possible. Of course, except in the rare natural occuring experimental situation, this is not possible for researchers on humans, unless the scientists worked in NAZI times, But I digress.
All sorts of factors in free living human populations come up, many of which relate to the outcomes of experiments. So called intermediate variables (confounders, effect modifiers, etc.), while we can try to "match" study subjects and controls in various study designs, this is far from perfect. So techniques were developed to gather such data, say smoking, or weight, and to try to adjust for their differential effects in generating results. With few such factors, simple tabular statistics, such as 2X2 table Chi square or T tests work, and with so called AVOVA techniques, (analysis of variance), even more statistical adjusting can be done.
But for many extraneous factors, and many thousands of study subjects, nore refined statistical techniques are needed. And these can also help to estimate the sizes and relevance of differences in study populations.so called "effect measures", and to guage statistical significance. As a group, these all are forms of what is called multiple regression techniques, and have direct impact on making predictions ("Models"), as in COVID -19.

Later, alligators. And I await careful reading and notification of my gramatical and other erors from my wanker detractors. Here, boys, come and get your treat. .