Click here to monitor SSC
SQLServerCentral is supported by Red Gate Software Ltd.
Log in  ::  Register  ::  Not logged in

Tim Mitchell

Tim Mitchell is a business intelligence consultant, author, trainer, and SQL Server MVP with over a decade of data management experience. He is a partner with Linchpin People and is the principal of Tyleris Data Solutions.

Tim has spoken at international and local events including the SQL PASS Summit, SQLBits, SQL Connections, along with dozens of tech fests, code camps, and SQL Saturday events. He is coauthor of the book SSIS Design Patterns, and is a contributing author on MVP Deep Dives 2.

You can visit his website and blog at or follow him on Twitter at

Electronic Health Records – What’s the Big Deal? (Part 2)

In the first installment of this discussion, I talked about the challenges facing database professionals and others with respect to healthcare data integration.  In this post I’ll talk about the first part of the problem: a lack of adoption of the required technologies and/or methodologies.  I’ll also describe when an electronic health record really isn’t.

docWe’ve all seen it; it is still the de facto standard of medical documentation.  You sit in the reception area and the nurse or clerk gathers the initial information: demographic information to confirm your identity, insurance information, the reason for your visit, and  metrics such as your blood pressure, pulse, or weight.  He also asks you about any medications you’re taking and any other physicians you are currently seeing, which you provide from memory as best you can.  This information is written down on your chart, a collection of dead trees that is then hung on the door.  You wait for a while and the doctor arrives, writing still more information on your ever-growing paper chart.  She writes your prescriptions on yet another piece of paper.  You’re then dismissed with a stack of papers – your prescriptions, a carbon of your hand-written visit record and diagnosis, a scribbled referral to a specialist, and a business-card size reminder for your next appointment.

Several concerns come to mind in this scenario.  If the patient forgets to remind the physician about another medication she’s taking, the doctor fails to note a stated medical condition, if a portion of the paper record is lost or destroyed (or simply illegible), or if a required piece of information is not collected, the system can break down.  Further, recording information in this manner will restrict significantly the ability to index, search, aggregate, or mine the data.  What if the provider wants to find out how many patients required hospitalization after presenting with a particular symptom?  Perhaps ER staff need to know if an unconscious patient has a particular allergy or medical condition?  Somebody’s going to be surfing through paperwork to answer these questions.  The most frightening part about that is that those answers are often needed immediately, and no dead-tree storage system can provide that.

What’s The Problem?

The problem is not a technological one, nor is it new.  So why does it still exist? Simply put, this archaic way of storing and accessing healthcare records usually works, and it has for decades.  Patients usually receive the care they need, providers almost always get paid, and most of the time the records workflow does not negatively impact patient safety. 

Why Change?

I’ve encountered a small number of physicians and other healthcare professionals who prefer to keep things the way they are.  Paper records work fine, so why change?  On the other side of the fence are those who see value in collecting and analyzing patient data, individually and in the aggregate, to improve patient care and overall process flow.  Support of EHR implementation is especially prevalent in the under-40 crowd, though there are exceptions that transcend every age demographic.

The chief arguments against implementing an EHR system are that 1) it’s expensive, and 2) it won’t bring much additional value.  The first point is true; EHR systems don’t come cheaply, and usually require significant soft costs as well (staff retraining and system upkeep among them).   To the second argument, there is rarely an immediate return on investment, but the long-term return can be significant.  With a fully electronic record system, staffing needs may be reduced because sending a patient’s health records to another provider requires only a few clicks and keystrokes rather than the time-consuming manual retrieval of paper and film.  Billing and collections can be streamlined by eliminating multiple points of data entry, further reducing workload and the potential for human error.  Patient safety is enhanced by allowing the rapid analysis of a patient’s data when making treatment or medication decisions.

Electronic Health Records that really aren’t

In my experience, there are few shops that are fully reliant on paper.  Healthcare providers often use a hybrid approach, tracking information such as patient demographics and scheduling with a modern database application but retaining the paper-based workflow of documenting diagnoses, procedures, and prescriptions.  Others modify this workflow and scan in the paper records after the fact, sometimes calling this an EHR.  Paper-to-electronic scanning solves a few of the problems (storage space and portability among them), but other challenges still remain.  To research a patient’s treatment history, someone still has to place eyes on each document in the person’s record, and it’s almost impossible to aggregate this information.  Even though these records are stored in an electronic form, they do NOT represent a true EHR system.  Scanning of hand-written paper records is a half-step toward the solution, but doesn’t cure the problems at hand.

Government Intervention

Finally realizing the importance of electronic health records, the federal government has mandated and is in the process of defining rules which will govern EHR adoption and usage.  Although they won’t be forced to use a true EHR system, providers will see their payments from Medicare slashed significantly if they fail to comply.  Further, there are significant short-term financial incentives for providers who implement EHR systems by the 2011 deadline.

The Future

cashregIn the next 20 years, we’ll see the use of paper healthcare documentation come to an end; shops that work on paper will be as common as crank-handle cash registers.  Government regulations and consumer demand will bring on a level of data availability and transparency not currently available in healthcare data; patients will have immediate, on-demand access to their own health records, and more importantly, will have the ability to research aggregated healthcare data to help take charge of their own treatment and evaluate the quality of their providers.  Further, healthcare outlets will be required to share information with one another for the benefit of all patients (I’ll talk more about sharing data in the future).

To Be Continued

In my next post in this series, I’ll discuss the sharing of healthcare data, including the challenges it presents and the possibilities it brings along.


Posted by Steve Jones on 28 October 2009

There's another item. paper records are there is the power goes down, BSOD, etc. In quick situations, they are still there. There's also the fact that someone's handwriting, especially in a fast moving environment, provides a nice record.

I like the idea of EHR, especially so that I can keep copies of my records, but I also think we might need receipts of some sort that can verify what's written down.

Posted by jcrawf02 on 29 October 2009


"There's also the fact that someone's handwriting, especially in a fast moving environment, provides a nice record."

Steve, haven't read many of your doctor's charts, have you? I know it's a joke, but, it's not...

And why not just move to tablet PCs that capture the written elements as well in a pdf or similar form? verified down the line to ensure data capture is accurate?

Posted by jcrawf02 on 29 October 2009

Interested in your thoughts:

Should more effort be put into open source projects for EHR to combat the disparity driven by cost pointed out in the above?

Posted by WillC9999 on 2 November 2009

<rant> Power goes down?! Indeed it does but paper records go missing, sometime permanently (e.g. fire). And what if you fall in on holiday and end up in the distant hospital? There is also the potential for illegible scribbles to be misread. Power outages are rare and can be managed in critical settings. You could get your 'receipt' via your personal EHR audit trail. You could (should?) have access to this directly - via a secure web interface. I would rather trust my health to a robust and well thought out EHR than a stack of paper scribbles! </rant>

Posted by les barkhouse on 2 November 2009; your questions, and maybe curiosity, answered.  I am the DBA here.  Our company is already doing this, on a tablet PC, and is integrating with hospitals quite nicely.  We must be doing something right when our product is used in 60% of the Ambulances in England... we are based on the east coast of Canada, but have customers worldwide.

Posted by Charlie Whitcraft on 2 November 2009

Great series. It would be nice that when you leave the Dr's office, clinic, or hospital you took an electronic version of your MR with you, with a small viewalbe PC app. Maybe after it was updated it would be backup up to the WEB in some sort of protected EMR file service. Many problems the least of which what do you use for identification of the MR or patient. What codeing systems do you use for coding vearious classifications of drugs, symtoms, severity, tests, etc. Big chalenges but big returns!

Posted by jcwcok on 2 November 2009

Other than expensive, and no value added, there is another significant reason.  That is most of EHR (or EMR) is not really design from physician-perspective and too cumbersome.  To use a EHR, it certainly will increase tremendous of workload to physician as well as slow down the process.  Therefore, physician-frendly is another key factor for EHR's market.

Posted by Terry Sigrist on 2 November 2009

To address expense we cannot afford to leave the Medical Record in paper format.  How is an EHR (EMR) expensive when compared to say a 10yr study going forward VS the same study using current data from the past? (See H1N1 correlation as example) How is it expensive when serving patients and per the cost of the study i.e. time, money, quality of care?   How is paperless expensive when paper files catch on fire, or destroyed in floods or a facility loses the record?  What about portability to another facility per a move, vacation, or consulting?  A physician can manipulate drop down boxes and dictionaries much faster than writing while maintaining data integrity.  It is many times faster to search a database VS shuffling through paper in hopes of finding the correct documentation.

Posted by Ed Salva on 2 November 2009

What standard are we working toward?

Is there a common UI for a physician who works accross several facilities.

Also,  I agree with jcrawf02 about the divide.  All facilities are required to adopt EHR/EMR yet the stimulus funds were only provided to the acute care facilities.

Posted by brad.ashforth on 2 November 2009

I work with SQL Server 2005 (2008 next year) in the Healthcare industry. We rely extensively on electronic records, but I guess that would be expected in IT. However I have noted in this last year (or two?) that my PCP (or whatever he's called now) has implemented PCs in each exam room. The PCP/Doctor <b><i>and</i></b> assistant both use it instead of paper. I agree that if the power goes out (or the network crashes, etc) then paper would have been nice to have. But as a diabetic w/HBP (who has lost 60 lbs in the last 10 yrs) I'd prefer to have my doctor have the availability of my historical database right there.

Posted by Richard Farris on 2 November 2009

Unless I'm mistaken. The EHR mandate was originally 2010, and has been extended to 2015, with a 1% medicare penalty (reduction in reimbursements) for each year thereafter (i.e. 2% in year two, 3%, etc.)

Posted by jgalt47 on 2 November 2009

I agree that EHR facilitate longitudinal studies, but that is not neccessarily a benefit to the patient or to the individual doctor.

Both my PCP and my cardiologist have switched to electronis recording within the last two years, but neither of them offered to give me an electornic version of my records or told me how to have another facility access them if needed.

So far EHR seem to benefit beauracracies but not the end customer.

Posted by HMK on 2 November 2009

The only problem I have with EHR as they are used right now, is that a lot of doctors seem to have forgotten the patient...  I went to a doctor's appointment with my mother recently and the only time the doctor actually came over and looked at her was when he took her blood pressure and listened to her heart.  The rest of the appointment he sat looking at his computer screen facing a different direction, asking her questions and typing the information in.  There are a lot of subtle clues to a patient's condition that can be missed if the doctor isn't actually observing, but is typing into a computer instead.  

Posted by Erich on 2 November 2009

I am surprised to read that anyone with tech experience would question the ROI of an electronic record system. The electronic records vs. paper records debate has been resolved a long time ago. The medical industry is no different than any other in this regard.

There is a tremendous opportunity for all of us here to develop a user(doctor, nurse, patient)friendly EHR system. It will probably end up connecting to an open source system like is. Then we can all improve on the UI's.  

Posted by knechod on 2 November 2009

This ia a problem that is HARD to solve.  It's not due to lack of will-power, but the immensity of the problem. At one point in time, it was estimated that banking had 5000 concepts to model.  Healthcare had 500,000 concepts. There is no database model in the world that captures healthcare.  There are insufficient standards.  We have hundreds of years of practice to try to marry to a revolutionary shift in technology.  To get an idea of the scope of the problem, browse through http:/

It will get there, but it will take time.  The idea of Medicare penalties is nothing new, but short-sighted beyond belief.  For another perspective on the issue,

Posted by bnordberg on 2 November 2009

I work in healthcare and have been attempting to merge EHR data across several large institutions - it is terrible. Somewhere out there people have the concept that if you have an EHR it will eventually save money as you can stop duplicate labs ... This is nonsense, until we get better standards. Every hospital collects data in different ways. I have been able to merge things like labs or pharmacy data. But you always run into units of measurement disparities (they can be converted, but it will require some manual review...), different levels of granuality, different names, different dosages... But the biggest problem is the method of collection is so variable. For example we found some clinics do fecal occult blood (FOB) tests and do not send results to lab tables. So to compare FOB results was impossible. As well, much of the data is now going into huge text fields. What do we do with that? NLP might be an answer, but that is a back end solution to a front end problem. Are we ultimately just overwhelming physicians with text data? For example my wife's EHR is well over 500 pages of text when printed - for a 14 day hospital visit. Several key radiology impressions were overlooked as they were on page 2.5 of a far too verbose reading of her CT's. The problem is healthcare really does not yet know what to do with information in the EHR - and that means there is no standardization.

Posted by mdv3441 on 13 December 2009

I agree with bnordberg, I do ETL from many different Labs, one application comes to mind provided by the CDC that allows anything to be entered in important fields. e.g. Field "Complexion" Entry "Pregnant", and if that's not bad enough, guess how many ways 'Brown' Hair can be spelled?

Leave a Comment

Please register or log in to leave a comment.