Homeostasis is not a phenomenon that is simultaneously achievable or not
achievable as that would again imply acceptance that dichotomous
classification is an appropriate system of classifying biological phenomena.
Here you misunderstand the concepts outlined in my email and indeed may have
confused the issue with the concept of statistical dependence (i.e. where
one outcome is dependent on another which is important in calculating
conditional probabilities - on tossing a coin you can't get a tail and a
head simultaneously!). The use of the terms relative and absolute also imply
you are looking at relative risk of an event happening and absolute risk of
an event happening (i.e. attributable rates) which of course we are not
discussing; perhaps you were referring to relativity of events?

In epidemiological research the importance of a precise definition of the
phenomena under investigation is crucial to the quality of the research
process. Encyclopaedia Britannica's definition of homeostasis may provide a
'neat' description of what, in gross or overview terms, homeostasis
involves, however does little to aid scientific understanding of what is in
reality a complex series of biological processes. Furthermore Britannica's
definition does not easily allow scientific investigation, especially
hypothesis testing, as it is too unwieldy, imprecise and not objective.

It is essential to understand that most biological processes (such as
homeostasis) exist as a continuum; the main point in my earlier email. That
is for each separate biological process that contributes towards (in your
words) "a desirable state of relatively steady health" there is a range of
values, or events, that all cells, tissues and organs perform within and
most of these conform to a normal distribution.

Again blood pressure again provides a good illustrative example:

Blood pressure (systolic, diastolic or mean) confirms to a normal
distribution if plotted for a given population. Clinically we look for
particular values and conceptualise those values into whether a person has
disease or not i.e. hypo or hypertension (This is somewhat of a misnomer as
in reality what the clinician is conceptualising is whether the patients
disease requires treatment or not but that is another story….). So we may
say that a patient has disease (hypo- or hypertension) or has no disease (a
normal blood pressure) which is where the concept of the clinically
convenient dichotomous classification comes in. The threshold of a higher
than normal or lower than normal blood pressure may be estimated from the
distribution curve for that given population and tests for disease (in this
case measuring BP) should be subjected to sensitivity, specificity, positive
predictive value and negative predictive value analysis.

This is a somewhat crude although convenient method of clinically diagnosing
disease however is open to multiple errors. If the test is not sensitive
(probability of a positive BP test given the patient has true hypertension)
or specific (probability of a negative BP test given the patient has not got
hypertension) then there may be a lot of patients wrongly diagnosed via a
positive test although they have not actually got hypertension (false
positives) and via patients with a negative test who actually do have
disease (false negatives). Needless to say the implications of this are
horrendous but it regularly happens though we try to keep 'falses' to a
minimum when desigining and piloting a medical diagnostic test.

The second main issue with this approach is the threshold at which patients
are diagnosed hypo or hypertensive. If this is set at a particular exact
value what about patients with values just below or above that value. (e.g.
JNC 1997 standardised definitions classify Stage I systolic hypertension as
between 140-159mmHg but what about the patients at 138, 139 and 160, 161
etc?)

The logical conclusion must be that the whole range or continuum of possible
values should be considered in interpreting a biological phenomenon ALTHOUGH
it is always easier, quicker and probably more practical to conceptualise
either "disease needs treatment" or "no disease needs no treatment" despite
the many adverse consequences of this approach.

Alistair Grant
Institute of Public Health
University of Cambridge



>From: "Stanislav A. Korobov" <[log in to unmask]>
>Reply-To: PHYSIO - for physiotherapists in education and practice
>    <[log in to unmask]>
>To: [log in to unmask]
>Subject: Re: The Myth of Homeostasis
>Date: Sat, 13 Jan 2001 15:55:00 +0000
>
>In connection with the Alistair Grant's e-mail of 12 Jan 2001:
>
>
>As I understood, you mean that such phenomena as homeostasis are
>SIMULTANEOUSLY achievable and not achievable. If so, then, strictly saying,
>you are right. Indeed, for each separate MOMENT, it is probably impossible
>to say whether this phenomenon is yet existing or already gone. However we,
>health and biomedical professionals, usually deal with more long periods,
>and I need, for example, to know did my patient have RELATIVE 'homeostasis'
>YESTERDAY or didn't. And I do not need to find out has/had he/she the
>ABSOLUTE homeostasis today or whenever. I simply know that such an ideal is
>not achievable.
>
>Accordingly, I am forced to deem that homeostasis is a phenomenon of
>dichotomous classification system. I.e. that an organism either has
>appropriate relative constancy and stability or it has not them. What
>'third
>case' may I anticipate? That my patient is in 'homeostatic state' while
>he/she is not in 'homeostatic state'?! Philosophically attractive but
>psychiatrically suspicious... No, thanks. Tertium non datur.
>
>In this relation, Encyclopaedia Britannica's definition of homeostasis
>looks
>greatly important: 'any self-regulating process by which biological systems
>tend to maintain stability while adjusting to conditions that are optimal
>for survival'. Nota bene: PROCESS. I.e. this is something changing at the
>moment, and we, in principle, cannot surely provide it with label
>'homeostasis' or 'non-homeostasis'. We can only assess -- to some extent --
>an intensity, dynamics, and, maybe, tendencies of forthcoming changes of
>this process. In most cases, I think, we'll quite be satisfied with
>information on an extent of stability of this process!
>
>On the other hand, I am inclined to the opinion that even a dichotomous
>classification system is too... multiple for biological homeostasis!
>Indeed,
>what is that case when an organism 'doesn't achieve homeostasis'? I am
>hardly realizing this situation. Ability to self-maintaining and
>self-regulating is one of the essential properties of a biological system.
>Consequently, until an organism is still living, it is about homeostasis or
>is keen to homeostasis as close as possible. What is an alternative to
>homeostasis? It is the absence of homeostasis. I know only such examples of
>this alternative as decompensation, necrobiosis, and death. Thus,
>homeostasis is rather a phenomenon of 'monochotomous' classification
>system.
>That is, it do not require such classifying approaches, it exists anywhere
>where living organisms exist. And saying that 'homeostasis is not achieved'
>is equal to that a living organism is about not to live... (Especially
>taking into account that stability of homeostasis is definitely relative.)
>
>I think we should look at homeostasis as a desirable state of relatively
>steady health. At least, presence of ability to maintain homeostasis at
>minimally possible energetical expenses is indicating that all seems to be
>to rights in this organism.
>
>P.S. I agree that phrases like 'plenty of research and clinical findings'
>are absolutely unacceptable in serious discussions. It is better to do
>without findings, in general...
>
>Stanislav A. Korobov, MD, PhD
>Physician-Physiotherapist
>P.O.Box 7, Odessa, 65089, Ukraine
>[log in to unmask]

_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.