This is the third of a group of articles that concerns science in relation to commerce and society in the case of medicine. We’ve already spoken about empiricism as one of the three forms of knowledge, but we haven’t said very much about what it looks like. At heart it is gathering verifiable facts from the real world as it is. The facts we gather, one way or another are measurements. Its seeming simplicity belies the real difficulties with this common form of scientific knowledge, especially those that concern groups and whether or not, in the case of medicine, it can be without a specific hypothesis.
As in the other articles, here is an AI podcast made by NotebookLM.
THE INEVITABILITY OF SAMPLING
The world is very large, so we can’t gather facts from anything more than a tiny corner of it. In other words, we have to gather a sample and make our measurements there. But how do we do that? If we want the sample to represent anything more than itself, we have to choose it very carefully. Once we do that, we need to say, as best we can, what it does represent.
By its nature, medicine has to do with people, so our sample is a sample of people. Compared to the world, even a very large sample is tiny. So whatever we find will apply only to some part of the world, certainly not to all people in general. Central to defining that part of the world lies in the idea of groups – identifying the people in our sample in terms that apply generally and mark them as representing similar people beyond the sample.
THE NATURE OF GROUPS
At the very least, to be a group, the people in it have to share some kinds of traits that can be specified by one or more facts. They live in one town for example, and represent what people are like who live in that town. Or they all have some disease and represent people with that disease. The list is endless: they don’t have any disease and represent normal people; they are children, or old people, or somewhere in the middle, men or women or both. Furthermore, our sample may have several kinds of groups. For example we wish a sample to include the sick and the well who live in some particular town, or country, or continent.
You will have noticed already that when I speak about sharing some kinds of traits, I am really saying they share certain facts. We may not measure such things as age, or sex for that matter, but we obtain them. There usually are other facts that we need to make clear what kind of group or groups we’re working with, even before we make our research measurements.
Grouping facts may need to be quite numerous, in some cases, in order to define those people beyond our sample to whom the research applies. For example, when we are studying a disease that has multiple stages, or forms.
THE GROUPING MEASUREMENTS
In the drawing below I hint at the kinds of measures that medicine uses to characterize groups. Such things as age, whether ill or thought to be free of illness, sex and other enduring characteristics of the person, and then a basket of unspecified defining traits. They are unspecified here because utterly unique to every research venture. What is essential is that the details of the basket be specified clearly enough that another investigator can replicate a similar group in order to confirm a particular investigation.
Grouping measurements highlight the unintuitive idea that all facts in scientific research are measurements. For example, to classify a person within a group as ill or well, how is that a measurement? It is essentially a summary of facts each of which, individually, is a measurement. In gathering the sample, one makes multiple measurements of each individual who reports a lack of illness. Such measurements may include body temperature, blood pressure, blood measurements, organ measurements, fluid measurements, just as in the research itself. But these measurements are not the research, they are necessary to define grouping.
THE RESEARCH MEASUREMENTS
Once we have our group or groups well defined, we need to make the measurements we came for, the ones that arose from the imagined objective, the ones which if obtained have a good likelihood of improving health for some definable groups of people.
Very broadly speaking, measurements in humans cluster within certain systems of the body, certain organs, or the body fluids. For example, blood pressure and body temperature represent signals that arise from the cardiovascular and central nervous or immune systems of the body. In the illustration, I show organ measurements that relate to glucose regulation, the kidney, and the GI tract. For blood, I mention the cells and the small constituents such as ions, but there are vast numbers of blood measurements in medicine. Likewise in urine.
The measures illustrated are nothing more than prompts to suggest the kinds of things empirical research seeks in the case of medicine. The actual lists of measurements would be mammoth. There is the genome, for example, the proteome in blood or urine, the transcriptome – this latter potentially arising from any of the cell types in the body. There is the microbiome, of the gut or urinary tract.
And there are visualizations of the body and its organs that return troves of measurements related to function, anatomy, and numerous forms of illness.
WHAT DO WE DO WITH EMPIRICAL RESEARCH RESULTS?
Broadly speaking, medical empiricism concerns measurements within or between groups, and whether the measurements are made once or are multiple over time. In all cases, however attained, empirical data are the ground upon which we build theory and out of which we create invention. It is essentially to say what is there in the human world that could be important for medicine.
In the case of measurements made over time, one can infer the likelihood of causal relationships and, with some rigor, exclude them. The reasoning is much like that of a murder mystery. If she was not there before he was murdered, and she was not there at the time his body was found, she is not a good suspect. Obversely, she is a good candidate, but proof requires something more.
Whether measured once or over time, new facts obtained within a group presumably add to the facts that are reliably bound up with the group and therefore extend the range of information over which we can define the group in the future.
New sets of facts that reliably differ between two groups, much the same. The facts help to define each of the groups more reliably than before and offer the additional advantage of differentiating the one from the other in the future.
In both cases, in adding new facts the research offers the potential of new insights into disease and therefore new tests and treatments in the long run. Sometimes the new insights are into the nature of health, which extend our understanding of the human condition.
THE EMPIRICAL HYPOTHESIS
With this general background in mind, let’s take a closer look at the empirical hypothesis. In the case of medicine it must be that either within a particular group or in the differences between two different groups there are important facts, both new and of use to certain people. Whether or not the interpretation of the facts will rest upon simple description, changes over time, differences between groups, or even changes in differences between groups over time is not crucial here. What is crucial is that there be a hypothesis more or less couched in these terms.
Some of the modern empirical techniques, such as the various ‘omes’, can seem resistant to the formulation of hypotheses such as these. They can seem almost an invitation to just measure so as to know the distributions of proteins, or nucleic acids within and between various groups of people, in this or that organ, fluid, or cell. But for the most part such hypothesis free accumulation of information has no necessary connection to the primary aim of research in the case of medicine, which is to discover new knowledge of value to specifiable people. It is, perhaps, a more valid research custom in what one might call basic science, where results need not benefit any one group of people apart from scientists themselves.
FINAL WORDS ABOUT EMPIRICISM IN MEDICINE
Put up against the brilliancy of invention and the profundity of theory, empiricism can seem drab, if workmanlike. To some extent, that is not untrue. Even so, one needs to imagine the way in which measurements made within well-chosen groups can improve the health of people, and that can be a demanding and subtle matter. More importantly, empiricism is often undertaken as an interim objective with the hope of using the information to work towards that objective which will certainly benefit a particular group of people. It is often one rung on a ladder of objectives.
If you think about it broadly, empiricism is about what is there. It is about the real world of interest in medicine, which is the condition of people, sick or well. It is as ubiquitous as water, and equally essential.