Free-text and narrative medicine
Khan, Alim
Dear all,
As well as being one of the facilitators for our upcoming meeting, I'd also like share a particular topic that I'm interested in exploring. It may particularly appeal to those oriented towards providing / improving patient care.
I'm interested in the (proper) use of free-text fields in EMRs - do you have these, do you use them, for what, how are they beneficial? Do clinicians use them for notes that don't fit in to other fields? Or are such notes recorded separately from other data in the EMR?
Why I'm interested -- I've been working with methods for aggregating large numbers of observations (unstructured data in narrative note form) and doing analysis to discover meaningful patterns in such data. One application of such methods we're exploring is for surveillance; others may be to discover other important measures that we might wish to track in a more structured way.
One part of this is the "narrative", the experience of a patient encounter, and what and how it is recorded. We're going beyond biomedical data to other observations overall. The niche field of narrative medicine suggests that clinicians can help themselves and their patients make sense of the experience of illness through constructing stories. I.e. the experience of living with HIV/AIDS is not just of interest in counselling or community health outreach but influences clinical encounters as well. To get an idea of narrative medicine, if you haven't come across it before, I'm copying an illustrative article from the Annals of Internal Medicine. Yes, emergency care in a developing country, but the setting is incidental: observations and sense-making of these happen in clinical encounters as well, maybe there is value to tapping into them in a more rigorous and systematic way.
Importantly, note that rather than constructing stories as per narrative medicine, I'm proposing to leave notes in raw narrative form and also capture some metadata. Such data may just be a few words here and there, rather than fully formed stories as per below. What we're talking about is deriving a bit of meaning from each encounter, then adding up lots of encounters to get something coherent.
Your comments and reactions welcome, either here or in person at the workshop. If there is sufficient interest I may demo the analysis software.
kind regards,
Alim
--
Alim Khan
World Health Organization
Geneva, Switzerland
kh...@who.int
+41 22 791 2924
Ann Intern Med. 2003;139:525.
Empty Pockets
Many of the experiences in our work we smooth and shape in the retelling like pebbles being polished by the tides. Some are more valuable for their sharp edges and the pain and lesson they bring by holding them tight. I spent part of my residency training at a city Veterans Hospital. Overnight call meant perpetual motion punctuated by frantic episodes of critical care. On this night we had done well, and by the time the first “code arrest” sounded overhead we all had managed to keep our heads above the tide. We arrived to the medical intensive care unit, wearing the authority our white coats gave us, to a bedside now ringed by staff. A trach collar had come loose and the trach itself unseated, but all had been corrected. Sorry about the run.
A curtain separated beds in this area of the intensive care unit, and each curtain had been pulled when the code was called. I lingered to one side, planning to stay long enough to see the collar retied. I was jostled by some movement at the adjacent bedside. The next time, it was followed by an arm. My name was shouted, and then I was pulled by my coat sleeve through the curtain to the next bed. The patient, a young woman of 36, had been admitted by one of the teams earlier that day. Chest pain, they said. The bedside monitor began to bleat and flash. Her blood pressure was falling with each systole, and her cardiac
tracing was taking on the jagged, wide, rugged contour that meant it would be a difficult trip. Her husband and two young daughters, 8 and 3, were led away from the bedside as the intubation tray was being
opened. Quickly intubated, compressions made obtaining central access more difficult, but I placed an internal jugular line hoping things would go well enough that it might be used for a Swan–Ganz catheter. Compressions, fluids, advanced cardiac life support protocols . . . all the while cardiac tracing into dysrhythmia. Occasionally, the monitor would only stare blankly asystole, across its screen. And the blood pressure. Boluses and drips and fluid, nothing would move the blood pressure. She was only 36. I reached deeper into algorithms, knowing the next vial off the code cart would run the answer into her collapsing circulation. The answer that would also push away the pain growing in my own stomach. I believe I tried everything, and trusted each new intervention would be the one. None of it ever changed the course. Boluses, drips, compressions, bagged breaths, time, and her heart finally stopped. She was only 36. I silently asked her why. A nurse touched my arm and said the two words whose weight made my next breath
difficult and whose grip was tight around the pain in my stomach. “The family.”
They had been taken to the waiting room. It was pale blue and the row of lights indicated the soda machine was still empty. My throat was dry. The youngest daughter sat on Dad’s lap looking at pictures in an outdoors magazine. The older sat watching her hands rest in her lap. Her husband’s eyes lifted to me and met mine. I didn’t, couldn’t, say a word. I took in the breath to begin, and he knew. The words never meant quite enough, “All we could . . .” They seemed small, like a string of toy boats against a stormy sea, “ . . . her heart . . .” I only had to say them, they would have to live them, “So sorry . . .” He turned back toward his daughters, a single father, and they lifted their eyes to his. As he drew the breath to begin, his eldest daughter knew. Suddenly all alone, she stood and ran toward us standing at the door. Dad bent toward her, but she got past him to me. Tears an unbroken line down her face, she took hold of my coat pocket. She said words that made 6 years of medical school, residency, and late nights too numerous wither. “Give mommy more medicine. She’ll get better, but you have to give her more medicine, please.” I couldn’t breathe. I had no words, no gesture. Her dad bent to her, but she wouldn’t let go of my pocket. He nodded to me, indicating I should go, and as I did the pocket she held tore. I drove home, tears making the road focus and unfocus, replaying it all to find the why or how. Ruptured ventricular wall? Aneurysm? Pulmonary embolus? I stopped when I realized it didn’t matter. I went home to my wife, then 33 years old. I took off my coat as she stood silently watching and put it in the trash. I needed some time to explain, all I could say was, “That one doesn’t work anymore.”
Connected Thinking
There are several interrelated issues—not for any one HCIT supplier or
region, but for all of them.
One issue is usually given a fancy name like “multi-vocality” or
“multi-valued predicates” or somesuch. Basically, multiple people
who are involved with the patient’s care are freely opining on various
things, and there is no assurance that all of them “agree” at any
point in time. Thus, some freetext narrative remarks may be
discordant with other freetext, or discordant with other coded values.
Additionally, any particular piece of data that is coded may be
annotated with freetext. The freetext in many cases is not a merely
descriptive detail but majorly alters the significance of the coded
item.
For example, a surgical explant of a VAD or a pacemaker or an ICD may
be performed because the device is no longer needed. Or,
alternatively, the explant may be performed because the device has
malfunctioned and MUST be replaced. There is no explants procedure
code difference to denote “explanted-for-cause” (ICD-9 37.64 or
37.97/98 or 37.85/86/87/89., respectively for the VAD, pacer, or ICD),
so the narrative text carries the main meaning as regards the “normal-
expected” vs. “abnormal-adverse-unexpected” nature of the explant
surgery.
Most of the currently-available NLP apps are not sufficiently powerful
to 100% disambiguate these—or other things like this example. At
most, the NLP apps these days are mostly serving as “helpers” to
ordinary human beings, who review and authenticate/resolve the
discrepancies or multivocality that the NLP apps discover.
http://www.ncbi.nlm.nih.gov/pubmed/17947628
Titus
mspohr
medical narrative illuminates the all-important 'soul' of medicine.
Unfortunately, this is difficult to capture in a digital world. If it
could be captured and represented, this would be a true breakthrough.
In the past all we have had are clumsy attempts to derive semantics
from free text and I think these can best be described as failures.
This is one reason that I have been a strong advocate of the minimal
data set. The minimal data set captures the small amount of
information that can be represented digitally. The remainder is in
the narrative and attempts to capture it digitally usually fail but at
great cost in time, effort, and complexity.
All the best for a new paradigm of 'mindfulness'.
Regards,
Mark
Mark H. Spohr, MD
Technical Officer - Health Knowledge Systems
Health Care Informatics
IER/HSI
World Health Organization
20, Avenue Appia
CH-1211 Geneva 27
SWITZERLAND
+41 22 791 24 04
--------------------------------------------------------------------------------
From: Ngeno, Titus [mailto:titus...@gmail.com]
Sent: 04 December 2008 15:55
To: Khan, Alim
Cc: Bailey, Christopher; Spohr, Mark; Whitaker, Patrick; Boucher,
Phillipe
Subject: RE: NLP and once-and-future 'narrative med'
Thank you, Alim. Attached another classic of medical narrative (this
one is “micro-nonfiction”), from JAMA 1988.
What most people in HCIT misunderstand is that the health care
services and processes are only nominally about
“medical” (physiologic) management. People in HCIT think they
understand what is needed but never are patient enough to stop and
think, to listen, to observe.
Yes, yes, quality, safety, cost-effectiveness. Of course. Duh.
But that is not all there is. Alim, you have happened upon some of
the artifacts of the lost civilization of clinical mindfulness, of
which I was once a full-fledged member. Today it is pure missionary
work. No one gets paid to be mindful anymore. In fact, people who
are explicitly un-mindful/un-caring are rewarded.
One small step for a person. One giant “Meh!” for humankind.
Looking forward to meeting you
Kind regards
Titus
------------------------------
From: Khan, Alim [mailto:Kh...@who.int]
Sent: Thursday, December 04, 2008 8:21 AM
To: Ngeno, Titus
Cc: Bailey, Christopher; Spohr, Mark; Whitaker, Patrick; Boucher,
Phillipe
Subject: RE: NLP and once-and-future 'narrative med'
Titus,
Very interesting perspectives. I do look forward to finding some time
at the workshop to engage with you on this topic. Briefly, I do see
that narrative medicine is coming from primarily the UK and US,
however the narrative traditions are universal, it is just a case of
finding how they are expressed in different settings. (And note that
there is a book for UK clinicians on how to incorporate narrative
medicine into 10-minute consultations). In most settings, however,
thoughts are not given voice or recorded in clinical encounters
(thinking of beyond a clinician's notes for their own purposes that
they may note on a record, to narrative observations). Here is where
free text in an EMR may play a role. Hadn't thought of the patient /
family member access but this makes perfect sense in light of the
whole social media / web 2.0 movement.
Now capture is one thing, and analysis and interpretation quite
another. Here I tend away from the interpretation of the text itself
(i.e. as in natural language processing, although great strides are
being made), to deriving significance from the relative positions of
different narratives to each other. Patterns / clusters are
meaningful, it's a question of on what basis do you group the
narratives. For this, some human-provided metadata is needed, i.e.
scoring along certain dimensions.
Anyway, this is a somewhat abstract discussion of a concrete method
(and software implementation) that we have used, we'll see if we have
the interest/time to do a demo.
FYI, attached is some info on a project from the Cognitive Edge
network, on global health issues, using the above methods.
kind regards,
Alim
--------------------------------------------------------------------------------
From: Ngeno, Titus [mailto:titus...@gmail.com]
Sent: 03 December 2008 15:32
To: Khan, Alim
Cc: msp...@who.int; Bailey, Christopher; Whitaker, Patrick
Subject: NLP and once-and-future 'narrative med'
Alim,
I am sure you know by now this topic is getting my attention! Just
tell me to stop..There is not much storytelling that is going to
happen in under 6 minutes… unless we have a taste for “microfiction”.
http://www.amazon.com/Micro-Fiction-Anthology-Really-Stories/dp/0393314324/
http://www.amazon.com/Flash-Fiction-Forward-Short-Stories/dp/0393328023/
http://en.wikipedia.org/wiki/Microfiction
Narrative Medicine has, you are right, been predominantly the province
of specialties (not just medical ones, but nursing and allied health/
social work ones) that deal with chronic and high-morbidity/high-
mortality conditions and disabilities. Encouraging the patients and
family members to annotate their EHR/CHR records or create diaries or
blogs and engage in Web 2.0 social networking is part of that. Same
thing for provider clinicians—who otherwise begin to exhibit features
of PTSD or other problems from the prolonged exposure to the stresses
of unsolvable situations.
Some of the best of the world’s literature and art has resulted from
just such storytelling. Here is the best of the sites, although
there are many others:
http://litmed.med.nyu.edu/Main?action=new
NLP of unstructured narrative entries can reveal many clues about the
implicit needs and perceptions and intentions that underlie what is
really going on. Sadly, this remains today a topic of research
interest only, since the cost-containment and risk-management
pressures that hold sway in the U.S. do not provide any way for
positive cashflow to be generated from creating or analyzing clinical
“stories”. And, in fact, patient advocacy and provider tort-law
experts tend to get very worried about the ramifications of
disclosures and privacy issues related to programmatic NLP of such
text chunks.
There is not much storytelling that is going to happen in under 6
minutes… unless we have a taste for “microfiction”.
http://www.amazon.com/Micro-Fiction-Anthology-Really-Stories/dp/0393314324/
http://www.amazon.com/Flash-Fiction-Forward-Short-Stories/dp/0393328023/
http://en.wikipedia.org/wiki/Microfiction
Narrative Medicine has, you are right, been predominantly the province
of specialties (not just medical ones, but nursing and allied health/
social work ones) that deal with chronic and high-morbidity/high-
mortality conditions and disabilities. Encouraging the patients and
family members to annotate their EHR/CHR records or create diaries or
blogs and engage in Web 2.0 social networking is part of that. Same
thing for provider clinicians—who otherwise begin to exhibit features
of PTSD or other problems from the prolonged exposure to the stresses
of unsolvable situations.
Some of the best of the world’s literature and art has resulted from
just such storytelling. Here is the best of the sites, although
there are many others:
http://litmed.med.nyu.edu/Main?action=new
NLP of unstructured narrative entries can reveal many clues about the
implicit needs and perceptions and intentions that underlie what is
really going on. Sadly, this remains today a topic of research
interest only, since the cost-containment and risk-management
pressures that hold sway in the U.S. do not provide any way for
positive cashflow to be generated from creating or analyzing clinical
“stories”. And, in fact, patient advocacy and provider tort-law
experts tend to get very worried about the ramifications of
disclosures and privacy issues related to programmatic NLP of such
text chunks.
Thanks for introducing this topic/opportunity
I will see you in person if all my Travel planning becomes fruitful
Kind regards
Titus
PS..I have been doing keen research on neuro-linguistic programming
Marc Mitchell
Marc Mitchell
Marc Mitchell, M.D., M.S.
President, D-tree International
www.d-tree.org
and
Lecturer on International Health
Department of Global and Population Health
Harvard School of Public Health
665 Huntington Ave
Boston, MA 02115
Room 1-1210
tel: (617) 432-6322
mmit...@hsph.harvard.edu
http://www.hsph.harvard.edu/faculty/MarcMitchell.html ( http://www.hsph.harvard.edu/ )
>>> mspohr <mhs...@gmail.com> 12/4/2008 10:08 AM >>>
What most people in HCIT misunderstand is that the health care
services and processes are only nominally about
think, to listen, to observe.
Yes, yes, quality, safety, cost-effectiveness. Of course. Duh.
But that is not all there is. Alim, you have happened upon some of
the artifacts of the lost civilization of clinical mindfulness, of
which I was once a full-fledged member. Today it is pure missionary
work. No one gets paid to be mindful anymore. In fact, people who
are explicitly un-mindful/un-caring are rewarded.
http://www.amazon.com/Micro-Fiction-Anthology-Really-Stories/dp/0393314324/
http://www.amazon.com/Flash-Fiction-Forward-Short-Stories/dp/0393328023/
http://en.wikipedia.org/wiki/Microfiction
Narrative Medicine has, you are right, been predominantly the province
of specialties (not just medical ones, but nursing and allied health/
social work ones) that deal with chronic and high-morbidity/high-
mortality conditions and disabilities. Encouraging the patients and
family members to annotate their EHR/CHR records or create diaries or
blogs and engage in Web 2.0 social networking is part of that. Same
of unsolvable situations.
there are many others:
http://litmed.med.nyu.edu/Main?action=new
NLP of unstructured narrative entries can reveal many clues about the
implicit needs and perceptions and intentions that underlie what is
really going on. Sadly, this remains today a topic of research
interest only, since the cost-containment and risk-management
pressures that hold sway in the U.S. do not provide any way for
positive cashflow to be generated from creating or analyzing clinical
disclosures and privacy issues related to programmatic NLP of such
text chunks.
There is not much storytelling that is going to happen in under 6
http://www.amazon.com/Micro-Fiction-Anthology-Really-Stories/dp/0393314324/
http://www.amazon.com/Flash-Fiction-Forward-Short-Stories/dp/0393328023/
http://en.wikipedia.org/wiki/Microfiction
Narrative Medicine has, you are right, been predominantly the province
of specialties (not just medical ones, but nursing and allied health/
social work ones) that deal with chronic and high-morbidity/high-
mortality conditions and disabilities. Encouraging the patients and
family members to annotate their EHR/CHR records or create diaries or
blogs and engage in Web 2.0 social networking is part of that. Same
of unsolvable situations.
there are many others:
http://litmed.med.nyu.edu/Main?action=new
NLP of unstructured narrative entries can reveal many clues about the
implicit needs and perceptions and intentions that underlie what is
really going on. Sadly, this remains today a topic of research
interest only, since the cost-containment and risk-management
pressures that hold sway in the U.S. do not provide any way for
positive cashflow to be generated from creating or analyzing clinical
disclosures and privacy issues related to programmatic NLP of such
text chunks.
Thanks for introducing this topic/opportunity
I will see you in person if all my Travel planning becomes fruitful
Kind regards
Titus
PS..I have been doing keen research on neuro-linguistic programming
On Dec 2, 6:12 pm, Connected Thinking <Titus.Ng...@gmail.com> wrote:
> Alim,
>
>
> *multi-valued predicates* or somesuch. Basically, multiple people
> who are involved with the patient*s care are freely opining on various
> things, and there is no assurance that all of them *agree* at any
> discordant with other freetext, or discordant with other coded values.
>
> Additionally, any particular piece of data that is coded may be
> annotated with freetext. The freetext in many cases is not a merely
> descriptive detail but majorly alters the significance of the coded
> item.
>
> For example, a surgical explant of a VAD or a pacemaker or an ICD may
> be performed because the device is no longer needed. Or,
> alternatively, the explant may be performed because the device has
> malfunctioned and MUST be replaced. There is no explants procedure
> expected* vs. *abnormal-adverse-unexpected* nature of the explant
>
> Most of the currently-available NLP apps are not sufficiently powerful
> most, the NLP apps these days are mostly serving as *helpers* to
Khan, Alim
I'm just back from a presentation yesterday evening on state-of-the-art
semantic analysis of text by one of the leading researchers in the
field. I think we'll all be suprised in just a few years at the effects
of the powerful new algorithms currently being developed. I would agree
however that current technology falls short, so we need not linger any
longer over automated semantic analysis.
But in terms of *capturing* narrative, it's precisely the technology
that has allowed large scale capture at minimal cost and time; there are
now numerous examples of narrative databases with tens of thousands of
entries, captured for the cost of setting up a web form on a server.
There are ongoing capture systems that flow into a set of indicators,
which turn out to be much less expensive than other ways of conducting
M&E. We're pushing the envelope though by thinking of this in a
clinical context.
Now we come to a fork in the narrative road: the narratives I'm talking
about are not fully formed stories as per the medical narrative examples
we've seen here. Yes, these are powerful expressions of the "soul" of
medicine. I'm talking more about the raw data: a few words or a
sentence or two, anything that might be significant or even notable in
the literal sense of the word. A classic example is Chris Bailey's "dog
bite" encounters (for those of you who aren't familiar with this "case
study", ask Chris about it at the workshop!).
Now if we have an analysis method based on aggregation, not
interpretation, we can find significance in the formation of clusters of
narratives according to certain criteria. As opposed to trying to
derive meaning from each narrative fragment in its own right through
semantic analysis. In effect we're adding up a little bit of sense in a
number of observations, to get something sensible (crossing over the
threshold of coherence into something meaningful). The key is to do it
right or you'll get nonsense.
What we're exposing here is the critical thinking and clinical reasoning
processes of expert medical practitioners, and thinking about how this
can be harnessed to produce data. Efforts in this direction have often
been decision-support systems e.g. that provide clinical pathways, or
alerts and reminders, etc. Of course these do not replace expertise but
complement it. Other ways to enhance expert performance are to
distribute the sensemaking/situation awareness/problem framing into the
environment. Most diagnostic aids fall into this category, everything
from lab tests to equipment used in a physical. To these we would be
adding the attention and awareness of those people in the room in the
clinical encounter.
Yes, you still need a minimum data set; beyond this, though, factors
that affect adherence or indeed any of the socially mediated factors in
HIV/AIDS treatment also affect quality of care. It's the messy
ambiguity of narrative that allows it to accommodate these kind of
factors, and the technology that can help preserve and even derive value
from the messiness. Filling in the gaps between the checkboxes on the
form, so to speak.
kind regards,
Alim
--
Alim Khan
World Health Organization
Geneva, Switzerland
kh...@who.int
+41 22 791 2924
Tim Cook
building knowledge models that are a maximal data set and then doing
constraints at the local level the complete semantics are maintained in
the data capture required in any specific application.
See:
http://www.openehr.org/shared-resources/getting_started/openehr_primer.html for an introduction to this approach.
Cheers,
Tim
> --~--~---------~--~----~------------~-------~--~----~
> You received this message because you are subscribed to the Google Groups "EMR Data Use Workshop" group.
> To post to this group, send email to emr-data-u...@googlegroups.com
> To unsubscribe from this group, send email to emr-data-use-wor...@googlegroups.com
> For more options, visit this group at http://groups.google.com/group/emr-data-use-workshop?hl=en
> -~----------~----~----~----~------~----~------~--~---
>
--
Timothy Cook, MSc
Health Informatics Research & Development Services
LinkedIn Profile:http://www.linkedin.com/in/timothywaynecook
Skype ID == timothy.cook
**************************************************************
*You may get my Public GPG key from popular keyservers or *
*from this link http://timothywayne.cook.googlepages.com/home*
**************************************************************
Connected Thinking
In my opinion, the Risk Management Dept and General Counsel of
healthcare institutions will be vigorous opponents of any attempts to
undertake a semantic processing to abstract and codify likely
intentions and “illocutionary acts” of the providers or patients who
generated the textchunks.
http://www.amazon.com/Illocutionary-Sentence-Meaning-William-Alston/dp/0801436699/
http://www.amazon.com/Diachronic-Pragmatics-Studies-Illocutionary-Development/dp/1556199465/
http://www.amazon.com/Foundations-Illocutionary-Logic-John-Searle/dp/0521263247/
http://www.amazon.com/Propositional-Structure-Illocutionary-Force-Contribution/dp/0674716159/
http://www.amazon.com/Minimal-Rationality-Bradford-Christopher-Cherniak/dp/0262530872/
Those lawyers and risk managers realize that, in the U.S. at least,
those sorts of information would be subject to ‘legal discovery’
process under applicable state and federal law. And there is
currently no benefit or advantage or positive cashflow that would be
produced by such NLP semantic processing. But, to the contrary, there
**would** be significant tort exposure and financial downside
generated by such processing if its outputs were discoverable by the
court and by attorneys-for-plaintiff.
So I think realistically the only way that the NLP processing for
programmatically inferring the implicit and illocutionary meanings
(the ‘soul’, as you put it) will ever get approval … is if the use-
case is a “quality-assurance”/”quality improvement”/”patient
ombudsman” one.
Alternatively, it may be possible to do it as an “observational/
translational research” use-case, provided that the records have all
the patient and provider HIPAA PHI information stripped off and none
of the patients or providers are identifiable.
Consider that under HIPAA rules, no one other than HIM/Medrec and CQI/
TQM/Risk staff have authorization to access individual medical records
without the assent of the patient/guardian or without an IRB-approved
research protocol. The caregivers who are (were) responsible for the
care of a particular patient can legally access that patient’s
information, but any other third-party person who was not involved
must either obtain explicit consent or be conducting authorized
research or be performing an authorized medrec or quality or risk
management role, with a written scope and purpose and timeframe.
In other words, generalized unsupervised NLP surveillance with no
written scope and no finite timeframe and no stated purpose more
precise than trolling to discover probable intentions is almost
certain to violate HIPAA regs. From the providers’ point of view, it
would resemble so-called warrantless wiretapping—disclosing aspects of
their confidential expressions in the EHR that are private and not now
accessible by anyone without prior assent and approval. The only way
to cure that is to confine and specify and control the use-case to
relate to HIPAA-sanctioned use-cases, such as formal health services
research studies or internal quality-assurance.
See you all soon
Regards
Titus
>
> >>> mspohr <mhsp...@gmail.com> 12/4/2008 10:08 AM >>>
> Sent: 03 December 2008 15:32
> To: Khan, Alim
> Cc: msp...@who.int; Bailey, Christopher; Whitaker, Patrick
> Subject: NLP and once-and-future 'narrative med'
>
> Alim,
>
> I am sure you know by now this topic is getting my attention! Just
> tell me to stop..There is not much storytelling that is going to
> happen in under 6 minutes* unless we have a taste for *microfiction*.
>
> http://www.amazon.com/Micro-Fiction-Anthology-Really-Stories/dp/03933143
> 24/
>
> http://www.amazon.com/Flash-Fiction-Forward-Short-Stories/dp/0393328023/
>
> http://en.wikipedia.org/wiki/Microfiction
>
> Narrative Medicine has, you are right, been predominantly the province
> of specialties (not just medical ones, but nursing and allied health/
> social work ones) that deal with chronic and high-morbidity/high-
> mortality conditions and disabilities. Encouraging the patients and
> family members to annotate their EHR/CHR records or create diaries or
> blogs and engage in Web 2.0 social networking is part of that. Same
> thing for provider clinicians*who otherwise begin to exhibit features
> of PTSD or other problems from the prolonged exposure to the stresses
> of unsolvable situations.
>
> Some of
>
>
> read more »- Hide quoted text -
>
> - Show quoted text -