symptom

[GoogleCodeExporter]

Barry Smith: Modify symptom definition to mention subjective experience of
the patient.

Original issue reported on code.google.com by albertgo...@gmail.com on 18 Aug 2009 at 11:54

[GoogleCodeExporter]

Barry Smith: I prefer to see 'symptom' (when specified in a disease-neutral 
way) to
refer only to those pathological processes which are experienced by the patient
subjectively (as they say); thus symptoms are: 
[experiences of] depression, nausea, pain, fatigue, anguish, etc. 
(and above all pain)
there is a broader use in the context of phrases like 'symptom of disease X', 
where
we would allow also objectively observable symptoms such as rashes or elevated
temperatures

Original comment by albertgo...@gmail.com on 9 Sep 2009 at 4:20

[GoogleCodeExporter]

make symptom a bodily *experience* of a patient

Original comment by albertgo...@gmail.com on 1 Dec 2009 at 3:22

[GoogleCodeExporter]

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.

Original comment by albertgo...@gmail.com on 5 Jan 2010 at 7:05

[GoogleCodeExporter]

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'

Original comment by sivaram....@gmail.com on 5 Jan 2010 at 7:32

[GoogleCodeExporter]

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.

Original comment by hoga...@gmail.com on 5 Jan 2010 at 8:54

[GoogleCodeExporter]

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.

Original comment by alanruttenberg@gmail.com on 6 Jan 2010 at 2:05

[GoogleCodeExporter]

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?

Original comment by albertgo...@gmail.com on 6 Jan 2010 at 3:47

[GoogleCodeExporter]

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?

Original comment by albertgo...@gmail.com on 6 Jan 2010 at 7:47

• Changed state: Started

[GoogleCodeExporter]

Bjoern Peters wrote:

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

Original comment by sivaram....@gmail.com on 7 Jan 2010 at 4:10

[GoogleCodeExporter]

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.'

Original comment by sivaram....@gmail.com on 7 Jan 2010 at 4:14

[GoogleCodeExporter]

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?

Original comment by hoga...@gmail.com on 7 Jan 2010 at 4:37

[GoogleCodeExporter]

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. 

Original comment by sivaram....@gmail.com on 7 Jan 2010 at 4:44

[GoogleCodeExporter]

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.  

Original comment by bjoern.p...@gmail.com on 7 Jan 2010 at 6:40

[GoogleCodeExporter]

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'?

Original comment by sivaram....@gmail.com on 7 Jan 2010 at 7:43

[GoogleCodeExporter]

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

Original comment by albertgo...@gmail.com on 5 Mar 2010 at 5:19

[GoogleCodeExporter]

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.

Original comment by albertgo...@gmail.com on 18 Nov 2010 at 8:27

• Changed state: Fixed

[GoogleCodeExporter]

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.

Original comment by alanruttenberg@gmail.com on 20 Nov 2010 at 10:08