Physician Answer Syndrome

I was recently reading a post about physician use of clinical decision support titled "Do Decision Support Tools Make Docs Look Dumb?"  What I find interesting about this is that I routinely Google it, or look it up on the web, or in a particular document.

The volume of information that I am expected to be conversant with is incredibly large and includes: Programming syntax details in at least 20 different programming languages, software library capabilities in over two dozen freely available or licenses libraries, the operation and automation of about a dozen different tools, how to administer two different database servers made by different vendors, how to administer a web server, three different UML modeling tools, how to administer an operating system, and how to implement about 100 different standards and profiles.  That's probably not as complicated as being a doctor, but it's still pretty complex.

I try to keep enough of that in my operating memory to do the day to day stuff, and to know where to go when it's no longer day-to-day.  When I don't know, I also know a bunch of others like me who I can ask.

I'm not ever embarrassed at having to look something up, but I'm also rarely ever put into a position of need by someone who has no understanding of my particular art.  Even on those occasions, I'm still not embarrassed.   It's fairly simple to explain that I needed to check something out, and to explain what I discovered.  The translation of the user's need into an appropriate query, and interpretation of the results of that query into something that the end-user can do is quite valuable.

There's an old joke I've heard a bunch of different ways applying to a repairman who is called to fix a problem.  He listens very carefully to the customer who gives quite a detailed description of the problem.  After about 10 minutes of listening, he says, I know what your problem is, and then goes over and makes a very simple adjustment.  His customer is very happy, until the repairman issues the bill.  It's $55 dollars for no more than 30 seconds of work.  The customer is outraged.  "This bill is way too much.  You cannot charge me $55 for less than a minute's work."  The repairman agrees.  "You're right.  Give me back the bill."  He scratches out the $55 and writes the customer a new bill and hands it back.
   Adjusting the Thingamabob:  $1.00
   Knowing that Adjusting the Thingamabob would fix it:  $54.00
The customer paid.

Another great story is physician related.  A doctor has an extremely difficult case.  He calls on a colleague who listens to him go on about the case.  After quite some time, his colleague motions to the doctor to wait, and then steps out of the room and returns with a book.  He reads the doctor the answer from the book, and then closes it and returns it to where he got it, returning back to the room where the doctor is still sitting, now dumbfounded and outraged.  "You are a farce!  You didn't know the answer, you simply read it to me from a book.  You are supposed to be the best in the field.  How can you do that?"  His colleague says "Follow me" and leads the doctor to a large room filled with books.  He waves at the room expansively and then turns to the doctor and says, "Now, which of these books holds the answer that you need?"

The point is, knowing how to solve the problem is what is important.  That includes being able to access the right information that helps you find the solution.  It shouldn't matter whether its a healthcare problem, a computer problem, or a car problem.

If providers find demonstrating that skill to be embarrassing, perhaps they may suffer from a variant of another disorder, perhaps we should call it "Physician Answer Syndrome".

Cervicular Radiculopathy: An unscientific experiment in HCSM and EBM

If you want to know what it is (and it isn't a funny looking anything), this is actually a pretty useful query.  As an Health IT expert and social intermediary I thought I'd use my recent diagnosis as an opportunity to see what patients who suffer from more severe diseases encounter when they look into disease information using the web and social media.

First of all, I applaud Google's efforts on putting NIH material at the top of the list of returns on my query.  That was most helpful and decided to use that content as a baseline against which I could compare other reading material.  Also, Google's spelling correction of my original spelling (Cervical Radiculapthathy) was great. Any search tool that doesn't include spelling corrections on complex medical terminology is NOT consumer friendly.  My first attempt to type in this term I didn't initially know how to spell produced an excellent result in Google because it spell corrected my mistakes.  But that was not true in any of three other medically focused search engines that I tried.  The Google results were much better organized than the medical search engines.

I tried Bing with the same query and got nearly similar results with TWO important distinctions.  One:  Bing didn't include the NIH material.  Two:  Bing totally failed my ad-hoc spelling correction test.  Come on Microsoft, I know you have spelling correction software ... use it.

As soon as I put the words "Blog" after my search criteria, my experience changed.  The quality of the material went down greatly.  Several blogs included copious Google Ads.  I hate sites that abuse Google AdWords with a passion.  That was enough to drive me away from any site containing healthcare data.  The site creator seems more worried about funding the site than the patient/user experience.  The other things that I saw were sites that were "single-treatment" focused.  These tended to be mono-focused on how "their product" produced results.  I shied away from these as I would a snake-oil salesman but I can see how this stuff is very scary to physicians.

A couple of blogs turned up with the HONcode logo.  That means they've made some effort to be visibly authoritative and conform to a particular code of behavior.  I did not find any other similar effort expressed in more than one site (nor would I have been able to recognize it).  I find myself giving those sites higher marks and it became something I actively looked for.  I used the HONcode search engine and found a rather wide variety of opinions on diagnosis and treatment of my particular condition.  I was frankly surprised somewhat by the variety there, especially given the general consistency I found within other sites I found with pretty authoritative content.

Most notable in many sites was a complete lack of citations to evidence-based-medicine.  The only site that got good marks for that was the NIH site.

A couple of things that I found to be indicative of good content in this brief survey of the web:

1. References to Evidence-based-medicine, especially NLM curated citations (e.g., PubMed).
2. Content attested by a well known clinical organization.
3. Content with a clearly identified last review date.
4. Professionally produced graphics and/or video (especially animation).  Content like that isn't cheap, and anyone who puts some effort into producing it is at least investing in the content, unlike the sites which used tons of adwords and short bits of text clearly lifted from online references.
5. Good indexing.  Often a medical term will have several significant variations (herniated disc was an alternate search term I used).  People who invest in making content accessible are paying attention.
6. Clear signs that the site conformance to a code of ethics (e.g., a link to ethics page, or logo like HONcode).
7. Good search technology.

And bad content:

1. Single treatment focused -- this drug, that therapy, et cetera, especially if the site is clearly benefiting someone who has a drug or therapy they sell.
2. AdWords... Especially at the top of the useful content ... nuf said.
3. Clearly copied content.

I looked separately for patient created content (the other side of the social media conversation).   A good quick test query for that was "I have diagnosis".  That sent me to a number of different message boards that I didn't recognize. Thinking about it a little, it seems obvious that these are pretty much wide open web-sites/message boards that I really would rather avoid.  I want to have others have established at least some level of "membership" before I gave them a way to access  content I created about me and my disease (this post doesn't count -- if it was a more serious or embarrassing concern, I'd be much more circumspect).   So, how would I find a more personally but still social site.  

I went to one pretty well-known patient healthcare social site only to discover that it fails to work with my favorite web browser (Chrome) -NOR- with IE behind the corporate firewall.  So I had to switch back to the home computer to access using my least favorite browser (I will note that it also worked on my iPad just fine).  Using my own equipment is more appropriate, but also raises some interesting policy issues for corporate America. For patients living with chronic health conditions, using employer equipment to access HCSM websites that help them manage their health may be quite worthwhile.

The Chrome failure made me wonder about the tech-savvyness of ePatients.  I expect most (but not all) patients using HCSM would be more tech savvy that the prototypical e-mail-using facebook-posting grandmother (that would be my mother here).  Hmm, I just equated ePatients with users of HCSM.  Is that a fair equation?  Probably not.  There are surely differences here, which requires more in-depth exploration.  Just because all the ePatients I know are also users of social media does not make them the same group.  And my mother, while not a techy-genious practically invented the PHR [she just uses Microsoft Word and a lot of paper to carry it around].

My search from that one social site turned up little for my specific condition without joining.  It did turn up enough aggregate anonymous data to let me know that I was already much better off than others who had turned to it (simply by reading the treatments used).  I didn't fit the typical patient profile (not surprising, considering this was an experiment encouraged by a real life event rather than real life use).  I had similar issues with bad search indexes and lack of spelling correction there.  They didn't cover Cervicular Radiculopathy in their search index, so I fell back to "herniated disc", which did work.  I also tried "herneated disc" and "herniated disk" which failed to turn up anything.  The average patient wouldn't have been as persistent as I was, but then again, the average patient probably wouldn't be at that page to begin with.   

As a patient, did I find my very brief foray into HCSM valuable?  Yes, I did.  Several of the blogs referenced therapy options that neither my PCP nor my PT mentioned.  I'll be asking about one of those options later this week, since I have some indications based on my PT assessment that it may be helpful.   I also discovered some information that about some early symptoms of Carpal Tunnel that my doctor reported to me, but which don't line up with what I've read myself.  I'll be asking him about them on my follow up visit next week.  I'll report my results with my providers after I follow up.  Would I use a patient driven HCSM site?  I think it would depend on my condition and needs.  Certainly I already find a great deal of support in social media for other activities both professional and personal.  I would certainly look for and be happy to find an online community to support my health needs were they much greater than at present.  Even now my social network provides a good deal of support.

This was a fairly safe diagnosis to be testing these waters with.  There are certainly some easy lessons to be learned in this particular space.  I wonder about the convergence between #HCSM and #HealthIT and #CDS and #EBM.  Some folks are onto an idea that would readily bring these together, but that idea is just starting to emerge.

My own prognosis is good, and while my current pain makes it difficult to work, treatment seems to be helping.  Interestingly enough... typing doesn't aggravate the problem, but writing code or XML does.  It has to do with how I need to use my body for the latter task.  Writing this blog keeps my hands fairly well in the same place and mostly involves small hand and wrist motions.  Coding or writing XML involves mouse + keyboard activity which uses shoulder, arm and neck (to track the mouse across the screen).  That is much more difficult for long periods of time.

Already I've redesigned my workplace a bit, getting a larger monitor and raising it up on a swing arm to counteract bad posture brought on by my aging eyes (and I also made my long overdue appointment for an eye exam).  I've readjusted my chair to ensure a more upright posture while sitting and that also helps.

Today I even got close to writing some XML... tomorrow I'll work on it more and maybe even test out coding again...

Book Review: Information Retrieval: A Health and Biomedical Perspective

I just finished my first reading of the Third Edition of William Hersh's Information Retrieval: A Health and Biomedical Perspective (Health Informatics).  It's quite dense so I will have to go back through it again several times.  If you want to understand why computers, the web, and search engine technology haven't yet made evidence based medicine ubiquitous, you should read this text.

Bill doesn't spend much time explaining the mechanics of how to implement information retrieval systems.  For that, my well-thumbed favorites include Information Retrieval: Data Structures and Algorithms and Managing Gigabytes: Compressing and Indexing Documents and Images, Second Edition.  He mentions the former in chapter 1 along with several other texts.

What Bill does explain quite well are the challenges of Information Retrieval as it relates to the practice of medicine and medical research, and vice versa.  As I said, this was, even to one who is well versed in IR topics (having worked in IR and linguistics for 10 years), quite dense content, and full of usefull information and references.  Quite a bit of the book focuses on the work of the National Library of Medicine, including PubMed and MEDLINE.

What I found most disappointing in the book was the rather scarce coverage of Information Retrieval as it applies to the electronic health record, but I should not be surprised.  That lack is not Bill's fault.  It is up to our generation to apply IR technology to the EHR, just as it was up to prior generations to apply them to electronic text and the web.  Chapter 9 delves into some of the issues of Information Extraction from medical records, something I spent about 4 years working on in a prior job.  Coverage is a little bit thin in this area, but then again, some of the products that do this today in a very small way (tagging utterances in text) are only just now emerging into the medical marketplace, nearly a decade after I started working on them.

Bill is very well versed in his topic.  Lotka's law (see page 49 of his book) seems to hold if you look at the rather extensive (64 pages) of references that he includes in the back.  While I would certainly expect some bias towards his own publications, only Anonymous seems to be more prolific in this space.

Bill also writes a blog.  You can find him at Informatics Professor.  I usually find his posts to be worth reading and tweeting about.

   - Keith

P.S.  In the interest of full disclosure, Bill gave me this copy, just as I will be giving him a copy of my book.

Expressions in HL7 Data Types R2 and Computable Clinical Guidelines

The HL7 Structured Documents Workgroup met with Clinical Decision Support today to discuss some of the issues with the HL7 Quality Measurement Format that would need to be addressed in the next release.  Bob (Dolin) gave a quick update on the NQF status.  Apparently the Meaningful Use measures have all been converted to HQMF, including all the value sets in all the Meaningful Use specified vocabularies (ICD-9-CM and SNOMED CT) and are in CMS hands.  We heard that there may be some sort of comment / vetting process as a later phase.

The issue that Bob wanted to address is the way to represent an expression in a computational language in a measure.  HL7 Data types release 2 includes the EXPR data type.  This data type is an extension of a data type of type T which has one new component: expression.  An example representation of this is shown below:

‹value xsi:type='EXPR_INT'›
  ‹expression mediatype='application/javascript'›
    foo.value.value - bar.value.value
  ‹/expression›
‹/value›

Now, by itself, this isn't completely useful, but when you put it inside an Observation that you are defining, the expression can be used to define how the value is computed.  There are a couple of other things that you need.  One of these is a binding from the variables foo and bar above to specific classes.

The HL7 RIM has a way to create bindings for the derivationExpr component of the act class, but hasn't defined how to create bindings for EXPR_T.  I'd stick with using the same mechanism for derivationExpr.  What could be done is something like the following:

‹observation moodCode='DEF'›
  ‹value xsi:type='EXPR_INT'›  ‹actRelationship typeCode='DRIV'›

    ‹expression mediatype='application/javascript'›
      foo.value.value - bar.value.value
    ‹/expression›
  ‹/value›


    ‹localVariableName›foo‹localVariableName›
    ‹observation›
       ‹value value='1'›

    ‹/observation›
  ‹/actRelationship›
  ‹actRelationship typeCode='DRIV'›

    ‹localVariableName›foo‹localVariableName›
    ‹observation›
      ‹value value='2'›

    ‹/observation›
  ‹/actRelationship›
‹/observation›

What this essentially says is that the outermost definition of the observation has a value.  That value is computed from information contained in two other named classes: foo and bar.  These classes are then defined to be local variables representing the named observation classes. 

So, why is this cool?  Well, it's something only a geek could love.  What it does is provide a mechanism whereby we can bind an HL7 class represented in XML to a programming language like javascript (Bob wanted to use GELLO, but I can hand you a book today on javascript if you really need it.  And you can probably already figure out how to access the classes.

The next piece of this is that it allows certain computations to be defined based on the contents of other stuff.

What is missing from this is the binding rules that tell us how to evaluate the named portion of the expression.  I cheated by using the binding rules for derivationExpr, which are very simple.  Those rules state that the named variables are contained within in derived acts.  I could have used other binding rules, e.g., that the named variables are contained within some other set of named variables.

What I like about this is that it gives me the missing pieces needed to define a Structured Document for a Clinical Guideline.  Those two pieces are what I call level 3 and level 4 of clinical guidelines.

The structured clinical guideline in my head has four levels.  Level 1 contains a header comprised of metadata used to allow the guideline to be found in a repository of guidelines and human readable content as an attachment, e.g., PDF or XHTML.  Level 2 contains the information structured into sections where each section is addressible, coded, and has additional metadata describing it, along with human readable content in a format like XHTML.  Level 3 are the definitions of things that need to be tracked to manage the guideline (e.g., heart rate, ejection fraction ratio, blood preassure, comorbidities, et cetera).  Level 4 is a way to bind to ANY computational langauge, such as javascript (my preferred language for reasons of reduciong complexity).

So we never did get to discuss the idea of how to build this thing in clinical decision support like I had wanted to, because we could hardly get away from the discussion about how what Bob wanted was already in scope of VMR [sort of like swatting flies with a sledgehammer].  But now I know the pieces are there.  It's time to start thinking more about how to put this together.

And see, I don't even need to worry about the Gello, Arden, GLIF, ... debate because any mediaType will do as the computable language.  The standard need not state a preference.

So, it looks like there might be enough to define a quality process that has measurement built in.  One of these days, I'm just gonna have to take a class on that six-sigma thingy-ma-bob.

G'Night all.

   Keith

P.S.  They tell me that there's a foot of snow back home.  I hope you all are enjoying it.

Top O' the Week

Top of the week is The Healthcare Standards Interconnections, a visualization of how the Healthcare Standards community works together.

 

The top three posts of the month are:

1. I wanna be an ePatient A rhythm track is being written (or so I hear), and Dave and I are going to YouTube this, looking for willing performers in Boston area..., maybe in Cambridge for the HL7 Working Group meeting?  In case you haven't registered for that, today is the last day to get the  Early Bird Registration rates.
2. Meaningful Use Standards Summary  There are a couple of other things happening with quality measures that I need to follow up on.  See for example this quick tweet from @drtonyah on smoking status, to which I owe a deeper response (Thanks for digging that one up).
3. MeaningfulUse IG for Public Health Surveillance likely to Change  I've heard no news on the last post so I owe a follow up on the Public Health Surveillance guide.  I know at least one meeting has already occured.

Upcoming Events:

Well, I already mentioned the HL7 Cambridge Working group meeting, but what I didn't say was that there will be an opportunity to attend a half day seminar on standards for meaningful use on Monday October 4th. 

 

Today is also the last day for IHE PCC Proposal Submissions for 2011 development.

 

Recent Events:

Yesterday I spent the afternoon at a Knowledge Management roundtable discussion on EHR adoption.  If you missed it, search #KMForum for a quick recap.  A few quotables:  Emminence-based medicine referring to providers who practice medicine based on their "authority", and "We are in our EHR puberty" referring to the fact that we are headed into a period of rapid change.  The topics ranged far and wide from the main point, and it was a very lively discussion.  I enjoy these meetings.  @janicemccallum who was also there will be writing a blog post on this.

 

On Tuesday I spent an hour with about 165 people from 8 different countries on a free webinar, describing the HL7 CDA and CCD Standards. There are some really good questions, but we didn't get time to address all of them.  Look back here next week for some of the answers.