Re: Issue 12 in ogms: symptom

[og...@googlecode.com]

Comment #2 on issue 12 by albertgoldfain: symptom
http://code.google.com/p/ogms/issues/detail?id=12

make symptom a bodily *experience* of a patient

--
You received this message because you are listed in the owner
or CC fields of this issue, or because you starred this issue.
You may adjust your issue notification preferences at:
http://code.google.com/hosting/settings

[og...@googlecode.com]

Comment #3 on issue 12 by albertgoldfain: symptom
http://code.google.com/p/ogms/issues/detail?id=12

Barry's Suggestions:
Symptom (general sense)
A bodily feature of a patient that is observed by the patient and is
OF A TYPE THAT CAN BE hypothesized by a patient to be a realization
of a disease.

Symptom (strict sense - which I prefer, because it is ontologically
coherent)
An experience of an organism that is OF A TYPE THAT CAN BE
hypothesized to be a realization
of a disease.

[og...@googlecode.com]

Comment #4 on issue 12 by sivaram.arabandi: symptom
http://code.google.com/p/ogms/issues/detail?id=12

The later part of the definition - 'realization of a disease', may be too
restrictive. Although symptoms are often the result of a disease process,
there are
many common exceptions:
e.g. nausea & vomiting in pregnancy, headache as a result of listening to
a boring
lecture, shivering in response to a cold environment, ....
Perhaps, the definition may be broadened using
'realization of a CHANGE'

[og...@googlecode.com]

Comment #5 on issue 12 by hoganwr: symptom
http://code.google.com/p/ogms/issues/detail?id=12

Just about any type of experience can be hypothesized to be the realization
of a
disease. I don't understand adding "of a type" to the definition.

Yes, many experiences we have are unrelated to disease, but given issues
with
seizures and neurological disorders, I am not sure there are any types of
experience
that could not be hypothesized to be the realization of a disease.

For example, I distinctly remember a patient who complained of frequent,
unexplained
crying spells. Because she denied any depression or reason for being upset,
and
because she said she didn't know why she was crying, we ordered an EEG, and
sure
enough she had temporal lobe seizure activity.

So crying is of a type that can be hypothesized to be the realization of a
disease.
So is euphoria (bipolar disorder), sounds and sights (hallucinations), and
so on.

So I would lean towards a symptom being an experience that the patient
hypothesizes
as being the realization of a disease, as opposed to of a type that could be
hypothesized.

[og...@googlecode.com]

Comment #6 on issue 12 by alanruttenberg: symptom
http://code.google.com/p/ogms/issues/detail?id=12

I'm with Bill on this one. The current definition doesn't allow for a
doctor to say something like: No, that pain
isn't a symptom of shingles - you bruised yourself falling off the couch.

It also doesn't allow for symptoms unrelated to diseases, such as those
related to adverse effects from
treatments, e.g. pain due to injection of vaccine.

[og...@googlecode.com]

Comment #7 on issue 12 by albertgoldfain: symptom
http://code.google.com/p/ogms/issues/detail?id=12

I agree with these points as well. I actually think 'symptom reports' may
be easier
to put in an ontology than 'symptoms' (in the same way that IAO talks about
information content entities rather than information entities). Here are
some of my
preliminary thoughts from an email to Barry:

I believe symptom reports may actually be lower hanging fruit than
symptoms...but I
will get to that below

A partial symptom report has the form:

"s is a symptom experienced by e (over an interval t)"

and a full symptom report has the form:

"s is a symptom of d experienced by e (over an interval t)"

where s is a particular symptom (e.g., my nausea)
d is a particular disease (e.g., my influenza)
e is a particular experiencer (e.g., me)
and, optionally, t is a particular temporal interval.

I am supposing all of the following (but they are all tricky to axiomatize):

(1) s is in the range of symptoms that can be experienced by e in virtue of
the type
of organism e is (or is it better just to say 'in virtue of the anatomy and
physiology of e'?). For example, a human can experience dizziness as a
disorientation of vision while walking, but cannot experience a symptom of
bats like
disorientation of sonar during echolocation.

(2) if e is a human, then s can be lied about by e.

(3) if multiple types of diseases can have s as a symptom, then d is one of
those
types and e hypothesizes that d is one of those types.

(4) if a full symptom report is true, then e has disease d and the
experiencing of s
occurs during the realization of disease d. More formally, t is a
subinterval of the
temporal interval for the disease course of d.

(5) no experiencers e1 and e2 can share symptom instances (too trivial???)

I say that symptom reports are lower hanging fruit because symptoms are
just those
entities that expressed by s in true symptom reports (i.e., they really were
experienced and they really do indicate the disease). Getting symptom
reports right
is more relevant to groups and applications likely to use OGMS. FYI, I've
been
talking to some people from PatientsLikeMe.com in a different context, but
see their
symptom listing here:

http://www.patientslikeme.com/symptoms

I think they would be interested in a way of both classifying symptoms on
the basis
of patient reports and training patients in how to report their symptoms in
a
standard way.
--------------------
Some foundational questions involving symptoms:
(1) Do all diseases have multiple symptoms?
(2) Do any diseases have necessary symptoms (i.e., symptoms which, unless
experienced
by the patient, preclude the diagnosis of a particular disease)?
(3) What are the identity criteria for symptoms? (my guess: if s1 and s2
can both be
substituted into a full symptom report while holding d,e,and t fixed and
both reports
come out true, then s1 = s2.)
(4) What is the proper relationship between a particular cough, the feeling
of a
particular cough, and the throat soreness resulting from that particular
cough? (only
the feeling of the cough and the experience of throat soreness can be
symptoms in the
OGMS sense right?)

Thoughts? Feedback?

[og...@googlecode.com]

Updates:
Status: Started

Comment #8 on issue 12 by albertgoldfain: symptom
http://code.google.com/p/ogms/issues/detail?id=12

It was suggested in the meeting of 1/6/2010 that we add a term 'patient
symptom
report' with the following definition

patient symptom report =def a communication from a patient about something
they
perceive as being abnormal about their body or life

suggestions on improving this definition are welcome.

could we then say that symptoms are just those things that patient symptom
reports
are about?

[Bjoern Peters]

I mostly with Barry's original definition.
Regarding Bill's comment: Not all instances of crying are a 'symptom', only those that the patient hypothesizes to be realizations of a disease. In your example, the woman complaining about her random crying spells they are a symptom, as she reported them knowing that they were not normal. Me crying after St. Pauli loose again in soccer doesn't become a symptom.

Regarding Alan's comment: The pain a patient reports is a symptom, independent of the doctor saying it is caused by another disease than the patient thinks. Symptoms should not be linked to a specific disease when reported by a patient, they should just be considered 'abnormal'. Agreed that there should be symptoms of adverse events and pregnancy etc. as well, which are not diseases.

Either we punt on other symptoms for now, and do:
symptom of disease=def:A bodily feature of a patient that is observed by the patient and is
hypothesized by the patient to be a realization of a disease.

Or more general

symptom=def:A bodily feature of a patient that is observed by the patient and is
hypothesized by the patient to be a abnormal for the current stage of his life course. .

As an aside: is 'bodily feature' a process? The current definition seems to imply that.

- Bjoern

----- og...@googlecode.com wrote:

> Comment #6 on issue 12 by alanruttenberg: symptom
> http://code.google.com/p/ogms/issues/detail?id=12
>
> I'm with Bill on this one. The current definition doesn't allow for a
>
> doctor to say something like: No, that pain
> isn't a symptom of shingles - you bruised yourself falling off the
> couch.
>
> It also doesn't allow for symptoms unrelated to diseases, such as
> those
> related to adverse effects from
> treatments, e.g. pain due to injection of vaccine.
>
>

> --
> You received this message because you are listed in the owner
> or CC fields of this issue, or because you starred this issue.
> You may adjust your issue notification preferences at:
> http://code.google.com/hosting/settings

--
Bjoern Peters
Assistant Member
La Jolla Institute for Allergy and Immunology
9420 Athena Circle
La Jolla, CA 92037, USA
Tel: 858/752-6914
Fax: 858/752-6987
http://www.liai.org/pages/faculty-peters

[Bjoern Peters]

The message below was send out several hours delayed by our email server which just recovered from an intrusion yesterday. Ignore the message - I wrote this before the call, and the suggested definition derived during the call matches exactly what I had in mind.

[og...@googlecode.com]

Comment #9 on issue 12 by sivaram.arabandi: symptom
http://code.google.com/p/ogms/issues/detail?id=12

Bjoern Peters wrote:

Or more general

- Bjoern

--

[og...@googlecode.com]

Comment #10 on issue 12 by sivaram.arabandi: symptom
http://code.google.com/p/ogms/issues/detail?id=12

This discussion is bringing out many of the subtleties associated with some
commonly used terms.
Regarding Bjoern's comment: It is not necessary for a patient
to 'hypothesize that something is a realization
of a disease' for it to be considered a 'symptom'.
- The woman with crying spells may not think that they are abnormal
and may not even report it (as is the
case in depression). Her family may notice a change in behavior and report
it. Now, the St. Pauli FC is a
different matter ….
- Patients with diabetes may not complain of pain due to an injury -
this is a symptom of diabetic
neuropathy.
- A patient with a mole may notice that it has grown but may not
think much of it. On a routine checkup,
the doctor may ask about the mole and patient may say 'oh yeah, I noticed
that it became larger during the
past year'. Here, even though the patient did not hypothesize that it is
the realization of a disease (a mole
turning malignant) or even that it was abnormal, it still is a symptom.

The usage of the terms 'symptom' and 'sign' is very broad and context
dependent, and for many examples it
just does not matter whether it is called one or the other. Given this
should we even try to define them so
precisely?

I am leaning towards a simplistic definition given in Wikipedia:
'A symptom can more simply be defined as any feature which is
noticed by the patient. A sign is noticed
by other people. It is not necessarily the nature of the sign or symptom
which defines it, but who observes it.'
It goes on to clarify that
'A feature might be sign or a symptom, or both, depending on the
observer(s). For example, a skin rash
may be noticed by either a healthcare professional as a sign, or by the
patient as a symptom. When it is
noticed by both, then the feature is both a sign and a symptom. Some
features, such as pain, can only be
symptoms, because they cannot be directly observed by other people. Other
features can only be signs, such
as a blood cell count measured in a medical laboratory.'

[og...@googlecode.com]

Comment #11 on issue 12 by hoganwr: symptom
http://code.google.com/p/ogms/issues/detail?id=12

I agree with Bjeorn. Not all instances of crying are symptoms.

My point was to raise the issue of whether any type of bodily feature or
experience
could NOT be in some circumstance or other, hypothesized to be a
realization of a
disease. And thus, the addition of "of a type" into the definition adds no
or
minimal information (and is a little confusing to boot).

With respect to Sivaram's last comment, what if the patient is a lab tech
who
performs his own blood count, sees that it is abnormal, and reports it to
his
physician? Symptom or sign or both?

[og...@googlecode.com]

Comment #12 on issue 12 by sivaram.arabandi: symptom
http://code.google.com/p/ogms/issues/detail?id=12

I agree 'not all instances of crying are symptoms'. My point was about the
part saying 'only those that the
patient hypothesizes to be realizations of a disease'. A patient does not
have to 'hypothesize' for a feature to be
called a symptom.

[og...@googlecode.com]

Comment #13 on issue 12 by bjoern.m.peters: symptom
http://code.google.com/p/ogms/issues/detail?id=12

Bill: Agree with your point that 'of a type that' does not add anything. I
had also
removed that part in my proposed definition, and it was also gone in
definition
decided on after the call (#8)

Sivaram: I thought we the examples you gave in which the patient does not
hypothesis
and report something to be abnormal but a doctor does to be signs, not
symptoms. So
there would be something like:

report of sign=def a in which a clinician/doctor/diagnoser documents
something they
perceive as being abnormal about a patients body or life

Once we have those two, we need no longer distinguish symptom vs. sign, but
rather
can focus on 'observable body or life abnormalities' such as 'growing
mole', and
figure out how to link them to realizations of specific diseases, e.g. that
some
'growing mole' are part_of realizations of cancer of the skin.

As for the edge cases between report of symptom and report of sign
mentioned before
(e.g. self diagnosis by the lab tech and report of behavior by family), we
can either
call them something other than sign/symptom reports, or subclass them, but
in either
case they would still be about observed 'abnormalities in body or life', so
can be
treated analogously.

[og...@googlecode.com]

Comment #14 on issue 12 by sivaram.arabandi: symptom
http://code.google.com/p/ogms/issues/detail?id=12

Bjoern: Patients do not come in telling all their symptoms. Instead they
only tell a few (called presenting
symptoms) and the rest are elicited during history taking. So, the (crying)
woman may have presented with
symptoms of headache. She may volunteer information on crying only after
being asked about behavioral
changes - she didn't think it was important to mention.

Bill: I would think that the lab tech. is conveying a lab test result -
therefore, it is not a symptom.

What I am saying here is that a 'symptom' is a patient experience - whether
the patient, the doctor, or
anybody else for that matter, thinks it is important or not. A 'sign' is
from an observer point of view. Some
symptoms can be verified (e.g. rash), but not others (headache) at least
not yet.

Perhaps some thing like the following may be sufficient:
Symptom = def. any feature which is noticed by the patient and
is 'generally' thought to be indicative of some
abnormality.
Sign = def. any feature which is noticed by an observer (clinician/nurse)
and is 'generally' thought to be
indicative of some abnormality.

BTW, what is the relation between 'sign' and 'report of sign'?

[Barry Smith]

>I agree 'not all instances of crying are symptoms'. My point was about the
>part saying 'only those that the
>patient hypothesizes to be realizations of a disease'. A patient does not
>have to 'hypothesize' for a feature to be
>called a symptom.

My proposed amended version of the definition of symptom is as follows:

Symptom (general sense)
A bodily feature of a patient that is observed by the patient and is

OF A TYPE THAT CAN BE hypothesized by a patient to be a realization
of a disease.

Symptom (strict sense - which I prefer, because it is ontologically coherent)

An experience of an organism that is OF A TYPE THAT CAN BE

hypothesized to be a realization
of a disease.

Examples: [experiences of] pain, nausea, depression, anxiety, fatigue ...

This revised definition allows it to be possible that babies (and
even some animals) can have pains -- because they can have
experiences which are OF A TYPE THAT CAN BE hypothesized by a patient

[Barry Smith]

>I agree 'not all instances of crying are symptoms'. My point was about the
>part saying 'only those that the
>patient hypothesizes to be realizations of a disease'. A patient does not
>have to 'hypothesize' for a feature to be
>called a symptom.

My proposed amended version of the definition of symptom is as follows:

Symptom (general sense)
A bodily feature of a patient that is observed by the patient and is

OF A TYPE THAT CAN BE hypothesized by a patient to be a realization
of a disease.

Symptom (strict sense - which I prefer, because it is ontologically coherent)

An experience of an organism that is OF A TYPE THAT CAN BE

hypothesized to be a realization
of a disease.

Examples: [experiences of] pain, nausea, depression, anxiety, fatigue ...

This revised definition allows it to be possible that babies (and
even some animals) can have pains -- because they can have
experiences which are OF A TYPE THAT CAN BE hypothesized by a patient

[og...@googlecode.com]

Comment #15 on issue 12 by albertgoldfain: symptom
http://code.google.com/p/ogms/issues/detail?id=12

Barry Smith's reply to comment 4 above:

> Comment 4 by
> <http://code.google.com/p/ogms/issues//u/sivaram.arabandi/>sivaram.arab
> andi,
> Jan 05, 2010

> The later part of the definition - 'realization of a disease', may be
> too restrictive. Although symptoms are often the result of a disease
> process, there are many common exceptions:
> e.g. nausea & vomiting in pregnancy, headache as a result of
> listening to a boring lecture, shivering in response to a cold
> environment, ....

these are precisely not symptoms!

> Perhaps, the definition may be broadened using
> 'realization of a CHANGE'

this would make the definition much too broad (e.g. you realize your
fingernails have
grown too long; you realize your cat is purring ...)

> Comment 5 by
> <http://code.google.com/p/ogms/issues//u/hoganwr/>hoganwr, Jan 05, 2010

> Just about any type of experience can be hypothesized to be the
> realization of a disease. I don't understand adding "of a type" to the
> definition.

I would like to propose the following definition:
Symptom =def. - An experience of a sentient organism that is of a type
instances of
which are hypothesized by their subjects to be indicative of disease.

By 'experience' here I mean something essentially subjective.

The point of the definition is that it points to an ontologically coherent
family of
phenomena. This is part of an effort to save the existing Symptom Ontology,
which is
currently a vegetable garden of signs, symptoms (in the defined sense) and
other things.

The reason for 'of a type' is to allow small infants and higher mammals to
have
symptoms, even though they cannot hypothesize -- the point is that they can
have
experiences of the same type as adults.

> Yes, many experiences we have are unrelated to disease, but given
> issues with seizures and neurological disorders, I am not sure there
> are any types of experience that could not be hypothesized to be the
> realization of a disease.

You go through life experiencing your surroundings through sound and color,
and so
forth. 99.9999% of those experiences are de facto such that they are not
hypothesized
as indicative of disease. That some type-similar experiences could be so
hypothesized
is not, I think, a problem. If scientists have a name for the relevant
experiences
and use it in describing patients we will add it to the ontology,
alongside 'fatigue,
pain, nausea ...'

> For example, I distinctly remember a patient who complained of
> frequent, unexplained crying spells.

These are signs -- the doctor could (in principle) perceive them; thus they
are not
subjective

> Because she denied any depression or reason for being upset, and
> because she said she didn't know why she was crying, we ordered an EEG,
> and sure enough she had temporal lobe seizure activity.

> So crying is of a type that can be hypothesized to be the realization
> of a disease.

It is not essentially subjective

> So is euphoria (bipolar disorder)

euphoria of the sort which satisfies the definition is, I think, trivially,
a symptom

> , sounds and sights (hallucinations), and so on.

Hallucinations hypothesized by their subject to be indicative of disease
are symptoms.

> So I would lean towards a symptom being an experience that the
> patient hypothesizes
> as being the realization of a disease, as opposed to of a type that could
> be
> hypothesized.

How do you deal with symptoms in small infants and higher mammals?

> patient symptom report =def a communication from a patient about

> something they
> perceive as being abnormal about their body or life

I think this would be bad to include without a good definition of
'symptom'. Also 'perceive' is wrong here.
BS

[og...@googlecode.com]

Updates:
Status: Fixed

Comment #16 on issue 12 by albertgoldfain: symptom
http://code.google.com/p/ogms/issues/detail?id=12

Some consensus reached at OGMS version 0.7 where symptoms are reinstated as
a defined class and the definition has been changed to

"A quality of a patient that is observed by the patient or a processual
entity experienced by the patient that is hypothesized by the patient to be

a realization of a disease."

Added comment: Symptoms are typically reported by a patient to their
healthcare provider.

Added comment: We can drop the quality-or-processual entity solution (thus
reinstating symptoms as universals) under a theory of purely
experiential/subjective symptoms. However this theory will go against
usage in practice and will require a good theory for mental experiences in
BFO.

This is a deeply philosophical issue about subjectivity (i.e., experiencing
a process in a first-person way is saying more than the relation
has_participant seems to say). Of more practical importance is the 'patient
symptom report' (OGMS:0000088) that iao:is_about the symptom. This term may
be more relevant for applications of OGMS.

The main focus is now on the relation between signs and symptoms and
diseases/disease courses, see
http://code.google.com/p/ogms/issues/detail?id=45.

[Barry Smith]

Albert
The definition is ambiguous because the scope of the 'or' is not clear.
Also it seems to imply that babies and animals cannot have symptoms
BS

[Albert Goldfain]

I will tweak the natural language def to address the scope of 'or'.  What is your proposal to deal with 'hypothesizing...' for babies and animals?

[Ludger Jansen]

This seems to imply that there are no unobserved symptoms.
Isn't it possible that symptoms are overlooked, ignored, etc.?
LJ

[Albert Goldfain]

That's an interesting consideration Ludger.  Especially since (if symptoms are experiences) it may be the case that I can only attend to the pain in my head and don't notice the less salient pain in my leg.  I am tempted to say that in order for a pain to be a symptom it has to be noticed...but I am curious what others think.

AG

[Sivaram Arabandi]

I am not sure what the problem is here:

>"This seems to imply that there are no unobserved symptoms."

If something is not felt/seen by the patient, then it doesn't exist. 

e.g. Patients with Diabetic neuropathy do not feel pain. Therefore, they end up hurting their hands and feet. In fact, the 'lack of pain' is a symptom of the condition here.

>"Isn't it possible that symptoms are overlooked, ignored, etc.?"

If a patient has symptoms and chooses to overlook or ignore - the symptoms are still there. Now, patients often don't tell everything at the time of presentation. A number of symptoms are elicited during history taking by the physician.

--Sivaram

[Barry Smith]

A quality of a patient that is observed by the patient or a process experienced by the patient that is of the type that could be hypothesized by the patient to be a realization of a disease.

- unfortunately there is still a problem in that only processes are realizations

I would in fact drop quality and just say:

A process experienced by the patient that is of the type that could be hypothesized by the patient to be a realization of a disease.

the symptom then is the experience of pain, the experience of anxiety, etc.

BS

[Ludger Jansen]

Well, if PAIN is not experienced, pain does not exist. But a patient might have red pustules on his back or behind his ears, where he himself can't see them. But that does certainly not interfere with their existence.

 

Suggestion: A symptom is any observable property of a patient that is a sign for a disease of that patient, where the observable property is caused by the disease.

 

The modifier "observable" might be redundant, for unobservables cannot possibly be sign.

 

Would that do?

Ludger

[Barry Smith]

On Fri, Nov 19, 2010 at 2:22 PM, Ludger Jansen <ludger...@uni-rostock.de> wrote:

Well, if PAIN is not experienced, pain does not exist. But a patient might have red pustules on his back or behind his ears, where he himself can't see them. But that does certainly not interfere with their existence.

The question is: are they symptoms, or merely just signs? 

The idea I am still pursuing is that 'symptom' means: something that can only be experienced by the patient.

Symptom is a subtype of sign; most signs being such that they can be experienced by e.g. doctor, family member

 

Suggestion: A symptom is any observable property of a patient that is a sign for a disease of that patient, where the observable property is caused by the disease.

 

we do not use 'property' (since it can mean so many different things)

also you are here presupposing that we have a definition of 'sign'

BS 

[Albert Goldfain]

Taking a step back for a second...to explain the original motivation to close this particular issue on the list...

The universals in OGMS (as any BFO-based ontology in the OBO foundry) are based on science.  Experiences are subjective, and thus, are private to the experiencer. The best science of subjective phenomena we currently have is one in which experiences are inferred from objective measurements/recordings of behaviour and language.  Therefore, the majority of our ontological efforts should go towards representing symptom reports (for adult human patients) and observations of behaviors indicative of symptoms (for human babies and animals). 

There is no consensus on principled usage of 'symptom' in even the tiny OGMS community, not to mention medicine at large.  However, the current OGMS definition of symptom is ontologically cohereent and will not break interoperability with any other ontology based on (objective) science since they will only ever need to refer to reports and observations about symptoms.

does anyone disagree with this?

-AG

[Barry Smith]

remind me what the current definition of 'symptom' is

and also, for the sake of comparison, how we are proposing to use 'sign'

BS

[Albert Goldfain]

Here they are. Both are currently defined classes since continuants
and occurrents are disjoint in BFO:

[Term]
id: OGMS:0000020
name: symptom
def: "A quality of a patient that is observed by the patient or a

processual entity experienced by the patient that is hypothesized by
the patient to be a realization of a disease."

[http://ontology.buffalo.edu/medo/Disease_and_Diagnosis.pdf]
is_a: BFO:0000076 ! defined class

comment: "Symptoms are typically reported by a patient to their

healthcare provider. We can drop the quality-or-processual entity

solution (thus reinstating symptoms as universals) under a theory of
purely experiential/subjective symptoms. However this theory will go
against usage in practice and will require a good theory for mental
experiences in BFO."

creation_date: 2010-11-18T11:02:10Z

[Term]
id: OGMS:0000024
name: sign
def: "A quality of a patient, a material entity that is part of a
patient, or a processual entity that a patient participates in that is
observed in a physical examination and is deemed by the clinician to
be of clinical significance."
[http://ontology.buffalo.edu/medo/Disease_and_Diagnosis.pdf]
is_a: BFO:0000076 ! defined class
comment: "Signs are typically reported/recorded by a patient's
healthcare provider."
created_by: agoldfain
creation_date: 2010-11-18T11:14:36Z

[Barry Smith]

On Fri, Nov 19, 2010 at 3:04 PM, Albert Goldfain <albertg...@gmail.com> wrote:

Taking a step back for a second...to explain the original motivation to close this particular issue on the list...

The universals in OGMS (as any BFO-based ontology in the OBO foundry) are based on science.  Experiences are subjective, and thus, are private to the experiencer. The best science of subjective phenomena we currently have is one in which experiences are inferred from objective measurements/recordings of behaviour and language.  Therefore, the majority of our ontological efforts should go towards representing symptom reports (for adult human patients) and observations of behaviors indicative of symptoms (for human babies and animals). 

There is no consensus on principled usage of 'symptom' in even the tiny OGMS community, not to mention medicine at large.  However, the current OGMS definition of symptom is ontologically cohereent and will not break interoperability with any other ontology based on (objective) science since they will only ever need to refer to reports and observations about symptoms.

 does anyone disagree with this?

seems that many web sources do disagree, e.g.:

http://www.medterms.com/script/main/art.asp?articlekey=5610

http://www.memidex.com/symptom+medicine

http://en.wikipedia.org/wiki/Symptom

symptom  (smtm) A subjective indication of a disorder or disease, such as pain, nausea or weakness. Symptoms may be accompanied by objective signs of disease such as abnormal laboratory test results or findings during a physical examination. Compare sign.

The American Heritage® Science Dictionary Copyright © 2005 by Houghton Mifflin Company. Published by Houghton Mifflin Company. All rights reserved.

[Ludger Jansen]

In any case, both Sign and Symptom should be children of Role.

 

Maybe two kinds of symptoms should distinguished:

(a) complaints or afflictions: symptoms subjectively experienced by the patient

(b) medical findings: symptons that can also be observed by people other than the patient, e.g. the doctor

 

This is inspired by http://de.wikipedia.org/wiki/Symptom

(The English version holds to the pure subjectivist version.)

 

I guess that the current discussion is focussed on (a). Interoperability problems can occurr if other ontologies do not exclude (b). They can also occurr from the use of "symptom" outside of the field of human medicine, like economics and finances, where there are no patients.

 

The definition of "sign" is especially parochial: There are plenty of signs, even in medical business, that are not dependent on patients. E.g., words uttered by nurses, blinks on screens, numbers on thermometers, ...

 

Best,

Ludger

 

 

----- Original Message -----

From: Barry Smith

To: ogms-d...@googlegroups.com

Sent: Friday, November 19, 2010 9:06 PM

Subject: Re: Issue 12 in ogms: symptom

[Barry Smith]

On Fri, Nov 19, 2010 at 3:30 PM, Ludger Jansen <ludger...@uni-rostock.de> wrote:

In any case, both Sign and Symptom should be children of Role.

This is clearly true for Sign. One advantage of my approach to Symptom, is that the latter would become a first-order universal in its own right; entities would be objectively symptoms, and not merely play a certain symptomy role 

 

Maybe two kinds of symptoms should distinguished:

(a) complaints or afflictions: symptoms subjectively experienced by the patient

'complaint' sounds like 'report'

 

(b) medical findings: symptons that can also be observed by people other than the patient, e.g. the doctor

 

I think this is what we mean by Sign

A finding involves something like recognition on the part of a doctor. and there can be findings with signs or symptoms, e.g. when the doctor recognizes/asserts that the patient is healthy

This is inspired by http://de.wikipedia.org/wiki/Symptom

(The English version holds to the pure subjectivist version.)

 

I guess that the current discussion is focussed on (a). Interoperability problems can occurr if other ontologies do not exclude (b).

We do not intend to exclude (b)

Indeed, given our desire to support objective science, it is more likely that we would exclude (a), and replace it by something like 'patient report'

 

They can also occurr from the use of "symptom" outside of the field of human medicine, like economics and finances, where there are no patients.

 

 

The definition of "sign" is especially parochial: There are plenty of signs, even in medical business, that are not dependent on patients. E.g., words uttered by nurses, blinks on screens, numbers on thermometers, ...

This is indeed a problem

Perhaps we should use 'clinical sign'

BS 

[Ludger Jansen]

If symptom is a subtype of sign, it there should not be any "objective" symptoms.

Whatever you have as subjective symptoms (pain, panic, etc) can also occurr and NOT be a symptom for a disease.

 

I agree that "complaint" sounds to much like "report of affliction".

 

Suggestion:

Role

- sign (sth that signifies sth for sb)

-- symbol (sign by convention)

-- icon (sign by similarity)

-- index or symptom (sign by causal relations)

--- medical symptom (symptom for a disease)

---- medical finding (medical symptom that can also be observed by persons other than the patient)

---- affliction (medical symptom that can only be subjectively experienced by the patient)

 

Best

[Barry Smith]

On Fri, Nov 19, 2010 at 3:53 PM, Ludger Jansen <ludger...@uni-rostock.de> wrote:

If symptom is a subtype of sign, it there should not be any "objective" symptoms.

Whatever you have as subjective symptoms (pain, panic, etc) can also occurr and NOT be a symptom for a disease.

yes you are right

my mistake 

 

I agree that "complaint" sounds to much like "report of affliction".

 

Suggestion:

Role

- sign (sth that signifies sth for sb)

-- symbol (sign by convention)

-- icon (sign by similarity)

-- index or symptom (sign by causal relations)

--- medical symptom (symptom for a disease)

---- medical finding (medical symptom that can also be observed by persons other than the patient)

---- affliction (medical symptom that can only be subjectively experienced by the patient)

 

this is way too ambitious for ogms

the top three terms would fall within the scope of IAO 

BS 

[Albert Goldfain]

I would love signs to be roles...but the problem is that bfo:roles are
borne only by independent continuants... so things like cough, which
everyone on the OGMS call wants to say is a sign, can't be a bfo:role.
Thus the disjunction (and defined class status) for signs. FYI, this
problem is compounded when we try to represent 'productive cough' etc.

These issues have been brought up time and again during OGMS calls.

-Albert

On Fri, Nov 19, 2010 at 3:53 PM, Ludger Jansen

[Sivaram Arabandi]

- 1

[og...@googlecode.com]

Comment #17 on issue 12 by alanruttenberg: symptom
http://code.google.com/p/ogms/issues/detail?id=12

I don't see this as a fix.
1) It's not necessary - I thought we settled on symptom reports.
2) It doesn't cover the case where a third party reports - family member
(google symptom reported by family member)
3) The use of the term is not consistent in the community

Symptom reports can be about anything that is hypothesized by the reporter
to be a consequence of a disease.

I don't know if we need to actually distinguish symptom reports from sign
reports.

The ontologically coherent class that Barry suggests should be added, but
not as "symptom". Let's say "experience" for now.

Symptom reports can certainly about them. If we were able to encode the
constraints properly we would say that a symptom report about an experience
must be reported by the patient who experienced it, or be a second hand
report that the patient said they experienced them.

[KerstinF]

> A *process* experienced by the patient that *is of the type that could
> be* hypothesized

> by the patient to be a realization of a disease.
>
> the symptom then is the experience of pain, the experience of anxiety, etc.
> BS

As a newbie on this list I struggle to catch up with all the in-depth
discussions. I just wanted to provide a pharma perspective that may be
related to "patient experienced": A core area today in drug
development are PRO - Patient Reported Outcome, Often combined with
Heath Economics: HEOR.

Almost all clinical drug development program include PRO / QoL
(Quality of Life) instruments /questionaries, see for example
http://www.proqolid.org/proqolid/search__1/pathology_disease?pty=1915
And in some disease areas the statistical analysis of QoL-dimensions
are are used as primary endpoints in clinical studies. Could be
reports also from relatives, teachers etc..

Some of the instruments on the linked page are developed of my boss.
Elof Dimenäs. He combines a PhD in QoL with Informatics and also leads
the regional HEOR organisation at AstraZeneca. If it could add any
value I think he would be happy to provide some more input on this
topic.

- Kerstin