Data Input: The final frontier in clinical computing

I used to say that mobility, or rather the lack of mobility, was a leading factor that has limited the widespread adoption of electronic medical records by physicians.  Indeed, until just a few years ago we really hadn't cracked the code on ubiquitous wireless connectivity, the mobile devices and form factors that truly enable clinical computing.  Healthcare providers are mobile information workers.  Solutions that chain doctors to desktop computers really don't meet the workflow needs of healthcare professionals providing services in busy ambulatory or hospital settings.

We've come a long way in the last few years and the story is getting better and better.  Powerful, light weight and long-lasting notebooks, Tablet PCs, Pocket PCs and Smartphones provide the needed horsepower and flexibility to meet the needs of most clinicians.  These devices are also starting to morph and merge into single computing machines that begin to look like a platform for managing all of your communication and collaboration needs.  An example of this trend can be seen in the DualCor; a very cool 5-inch Tablet PC / Windows Mobile 5.0 combination device that serves as both a cell phone and full-featured computer.  Microsoft's Office Communicator application running on a wireless Tablet PC or notebook computer also takes us in that direction since I can now manage telephone, messaging, e-mail, web conferencing, and even video conferencing from my mobile device as long as it is connected to the Internet.

All that is good and well, but we are still confronted by one very thorny issue when it comes to clinical computing, and that is data input.  It doesn't make economic sense to require or expect that our most highly educated and expensive workers should spend their time typing at a keyboard or laboriously entering vast amounts of data into dozens if not hundreds of fields on some form. 

One of the things I have observed when visiting countries renowned for their use of electronic medical records in healthcare is that physicians are often not the ones entering data into these systems.  It is still very much a world of docs barking orders at clerks who enter the data for them.  This may not be ideal, but it certainly works well for the physicians.  Here in America and elsewhere, this problem is usually solved with voice dictation and transcriptionists. It's an efficient, but expensive practice that often serves as a kind of gold standard for data input speed when clinicians compare what they know (dictation) with what they are expected to do when using many of today's electronic medical record systems.

All of this points to the need for much better ways to capture and document the vast amounts of data that swirl around the typical physician-patient encounter.  I sometimes think that we should just wire up our healthcare facilities with video cameras and microphones and digitally record everything that transpires, leaving doctors and nurses with nothing more to do than just take care of patients.  Of course, that idea would probably never get past the privacy police.  None-the-less, it is appealing.  I believe the reason physicians so much enjoy doing charity work in the third world is because they are allowed to take care of people who really need and appreciate their services without the associated hassle of exquisitely documenting everything than transpires.

So, how do we solve this dilemma?  Does the Tablet PC take us in the right direction with digital inking?  Is the holy grail of data entry computerized speech recognition?  Do we hire more clerks?  Do we hook up those video cameras? Do we continue to dictate all of our notes and patient encounters?

What do you think?  We'd like to know.

Bill Crounse, MD   Healthcare Industry Director   Microsoft Healthcare and Life Sciences   

Comments (16)
  1. Have you thought about how you might apply Kim Cameron’s ( proposed ‘laws of identity’ ( to this to square any privacy concerns?

    This problem isn’t just about health of course – there’s a wider issue of how technology can be used across a whole range of occupations to let the experts focus on what they’re best at, to reduce bureaucracy, improve the quality and timeliness of services, and ensure sufficient reliable records are maintained and accessible when required.

  2. The real short comings of these platforms is software. While TabletPC is a great platform, I am somewhat biased since I love mine, it lacks serious software support in medicine. I had an OQO ( to evaluate and what i realized is that the software is not meant to run on it. This particular device had the further problem of being to big to be a PDA and too small to be an effective tabletPC. Maybe the right software could of made it work but as it was I just threw it back in the box and sent it back.

    Alot of software players in medicine pay only lip service to workflow. I often consult with engineers and architects at comapnies who have never seen their product used. This is the real problem.

    I think almost any form factor can work but the software needs to support it and support it well. Since you bring up transcription, which as someone involved in medical imaging is near and dear to my heart, check out Commissure ( Their solution is the best Speech Rec/Structured reporting solution that I have seen.


  3. hlthblog says:

    Thanks for your comments, Steve. You might want to take a look at some of the very fine Tablets from Motion or Toshiba. Motion has a new compact machine, the LS800, that may interest you. I’m a big fan of their full-sized Tablets; the LS1400 and LS1600 series.

    While hardware is important, I agree that it is the software that brings it all together. Part of the reason I write this Blog is to encourage developers and ISVs to take greater advantage of the devices and software on the market today, and to draw attention to the critical importance of involving clinicians in the design of better software for the healthcare industry.

    Bill Crounse, MD

  4. Shawny Marsden says:

    Good day to all – I have a child receiving home health care and have often asked "Why is the HHC industry still so paper-based". It is easier to implement electronic records when you have a hospital or doctor office setting – but how can the same be implemented in the home health industry? I have REAMS of records from the years of healthcare that my daughter has already received – records that became mine when she switched agencies. I have asked agencies the question and they all say that there isn’t a solution that is affordable. The HHC market is huge and the recent statement that 1 in 17 new births results in a defect – would suggest that it will get bigger. So how can we address the needs of this industry?

  5. Funny you should bring up this topic. I was just working on a blog entry discussing this very thing.

    I recently watched the Tom Cruise movie Minority Report. Detective John Anderton, trying to prevent a murder that takes place in the future, stands in front of a massive translucent heads up display and orchestrates a rapid fire array of graphical and text data and follows this dizzying cascade of information to conclude exactly when, where, and how this murder is about to take place. Armed with the information he heads to the scene of the "future crime"

    in order to prevent the death from taking place.

    This is the perfect model for healthcare in the future isn’t it?

    To expound on this thread, check out my blog.

    Douglas Krell MD


  6. hlthblog says:


    Thanks for your comment. I totally agree on the need for better tools for our home health providers. They are the heros of our industry. Home health may be the only thing that saves our healthcare system when all the baby boomers start to crump. We’ll need every provider we can get, and the only way we’ll get them and keep them is with better pay and better technology to help them do their jobs. Good IT is available now. Devices like Tablet PCs, Pocket PCs and Smartphones and software developed specifically for the home care market (such as Misys Homecare) are making a difference. In fact, visiting nurse services is one of the strongest markets for the Tablet PC. Yes, we have a long way to go, but the future is brighter than what you might think.

    Bill Crounse, MD

  7. hlthblog says:


    I really appreciate you weighing in. We need more clinicians helping developers come up with the next generation of great solutions for healthcare providers; people who really understand clinical workflow and how physicians and nurses think and perform their duties. We work in one of the most intensive information-driven industries on earth. The decisions we make, based on the information we have, may be the difference between life and death. I have colleagues here at Microsoft and in our research group who spend all their time thinking about how to improve the user experience and how to advance the way we interact with our computers. I even had a dialogue with a radiologist recently who was interested in how game controllers might be used to input data in clinical systems. So, stay turned. Your Tom Cruise analogy may not be that far off.

    Bill Crounse, MD

  8. Charles Beauchamp MD, PhD says:

    Experience from the VA EMR and extrapolating thoughts:

    1) Form Filled Hypertext – team member, preferably NOT the physician, using outline questions to check off answers and in so doing fill the note with pertinent conclusions and the database with guideline coverage. For example, if there is the goal of screening for depression, significant concern about pain, risk of falls, past history of sexual abuse, patient understanding of the important of diet and exercise…..etc, there should be a Form Filled Hypertext capability as is in VA’s VistA and CMS’ Vista-Office.

    2) Annotation of the Problem List – primary care physician can annotate a problem on the Problem List with the essence of the dx, rx, pt ed concerning that problem. In the VA this enabled clinicians to quickly survey the records of the New Orleans VA veterans being seen and migrate from the Annotated Problem List to progress notes and medications on the Medication List. This, in turn, produced a follow-up scenario for the New Orleans VA patient that was FAR SUPERIOR to what the patient being seen from Charity Hospital and University Hospital. Couple an Annotated List (Problem, Medication, Allergy….etc) capabiity with voice dictation of annotation snippets along with Ray Ozzie’s SSE (simple shared extensions) connecting to a HIPPA certified national database of Annotated Lists so that there is a non-VA capability of doing this. This could also connect into a Smart Card that the patient carries with her/him to have their annotated lists available for review and further annotation, even after evacuation.

    3) Revision of the data storage "allowances" in the EMR to incluse the ability to store (with appropriate HIPPA level security) the voice dictation of the clnician so that "the complete note" is not transscribed but is stored as a sound file and available (excuse the MacVocabulary) as a podcastable entity. Couple this with the Form Filled Hypertext and Annotated Lists and one should have a full record available at a LOT less expendature of physician time, as long as the physician can organize her/him-self around capturing the essence of the visit in the form of a typed outline – the Annotated List.

    4) Modularize everything so that the physician receiving a "view alert" (VA terminiology or clinical alerts) concerning an abnormal lab value or a specific patient request, can repond to the "alert" by having an attached order form available to fill out. This means that order entry must be connected to the view alert system, something the VA now does in part by allowing the physician to click on the view alert and go directly into the patient’s EMR.

    5) Catalyze a "bottom-up" approach to developing connectivity between observations and specific tasks that have proven efficacy in improving outcomes relevant to the observation. In computer talk that means if one has, for example, an American American person with congestive heart failure the mere mention of CHF on the problem list evokes a link to an order for Bidil (the hydralazine-nitrate combination). The clinician does not have to carry out this order but is presented with the very latest in "evidence-based medicine" for improving outcomes in well defined cohorts. Microsoft should sponsor an ongoing "gathering" of primary care clinicians across the globe to enhance the proposal of links that make "translation" of research into practice more than a pipe dream. Imagine the impact this could have, for example, in the midst of a world wide "Bird Flu" epidemic where clinicians across the globe make observations regarding ways to decrease the "cytokine storm" that causes most of the deaths, even in young and healthy folks. A clinician in South Korea could come up with simple nutraceutical means of downregulating cytokine production in Bird Flu infected patients and through means such as SSE connected didactic content translate those observations to all clinicians struggling to contend with what is otherwise an overwhelming problem. Of course, one would have to sort out the snake oil salespeople from the true observers and extraplators but with telemedicine spanning the globe and appropriate human-based filters this would be feasible, especially if one has a system, such as the VA has, of tracking outcomes in relationship to defined interventions with a nationally (and internationally) rolled up database. Never, never doubt that there are 10,000 local Schweitzers out there with the guts and intelligence to share their successes "in the bush", if only given the chance. Catalyze the combination of experience, evidence and theory while allowing the primary care clinician to access the ordering of efficacious tasks with an ease "like rolling off a log".

  9. hlthblog says:


    Thanks so much for sharing your experience, perspective and ideas. I urge my clinical colleagues to step up and get involved. You know better than anyone else the unique workflow and documentation requirements of clinical practice. You are uniquely positioned to help the IT industry and developers come up with solutions that will work for our profession.

    Happy weekend, everyone. And enjoy Super Bowl XL (go HAWKS)!

    Bill Crounse, MD

  10. Charles Beauchamp MD, PhD says:

    The following is a list of components for a PCP-centric web application with the implied requirement that there are multiple independent and inter-related control flows for these components in the EMR.  Each of these components can have "data input" requirements.

    Database retrieval & analysis

    Annotated List data entry

    Didactic material retrieval

    Decision support



    Communicating & Collaborating



    Model building



    It is my opinon that, as much as possible, the PCP (primary care practitioner) should be expected to input data via the Annotated List route.

    Annotated Lists are Lists of key elements that have outline annotation.  In the Windows world, Ultra Recall ( is the best example of "outline annotation".

    A PCP-oriented list of annotatable elements:

    Barrier to Care







    Supportive Content

    Belief, Value, Attitude




    Present Illness


    Pertinent Review of Systems

    Review of Systems


    Outside Medication

    Allergy / Adverse Reaction


    Development Milestone

    Childhood Illness

    Life Event



    Medical, Surgical & Psychiatric History

    Past Medication History

    Social, Relationship & Sexual History

    Family History

    Substance Use & Addiction History

    Risk Factor

    Procedure History

    Complication History

    Past Abnormal Result

    View Alert

    Past Care Plan

    Physical Examination


    Multi-Disciplinary Evaluation


    Discharge Summaries


    Reasoning / Decision Support

    Rule / Algorithm / Protocol






    Result & Delta


    Care Alternative



    Intervention / Service / Task

    Multi-Disciplinary Process



    Follow-up Care Plan

    OPC-based Trial





    Continuing Medical Education




    Pertinent components of these lists to a visit of the patient with the PCP can be patient, PCP, team member and data-driven.  Therefore, there should be manual and automatic inference capabilities that recursion through these lists to derive "pertinent components" that themselves could require further data entry.

    The danger, of course, is that the PCP can be easily overwhelmed by implicit and explicit data entry requirements.  Beware the wrath of the PCP if driven to spend most of the visit time being dictated to by the computer.  This is progressively so in the VA where the number of requirements being placed on the PCP is escalating with time with the requirements being imposed by "central office", regional units (VISN’s), local facilities (Hospitals and Community-based Outpatient Clinics), Payers, JCAHO and other review bodies, and care teams.

    These become "unfunded mandates" that are progressively making primary care untenable, inside and outside the VA.  These "unfunded mandates" make sense to the individual imposers but they result in a "tyranny by the computer".

    There does have to be a PCP "command and control" element to data entry and a constant attention to the view of the PCP.

    We are entering an time when there will be a focus on using the EMR as a panacea to some of the ills of healthcare, including "medical error reduction" and "cost-effective, evidence-based care", all the while only a fraction of the data viewing, data entry and communication support issues have been thought through.

  11. In the UK 100% of Primary care Practices are computerised, and 75 have had computers on the clinicians desk for over a decade.  Small I.T.companies, often with doctors as directors, fought over a possible 10,000 contracts. As a result ergonomic software, dealing with complex issues, coding and prescribing, using hundreds of quick keys, were developed.  Speed of data entry in front of the patients, without loosing eye contact with the patient, was and is essential.  Mouse is movements and tabbing fields is not good for data entry in the clinic.

    These "legacy" systems collected and posted some of their  collected data the UK has the largest prevelance data base in the world

    One IT medical IT company had to shelve its first Windows interface product as Primary Care doctors and their staff rebelled at how slow getting data in was on such a platform.  

    The disadvantage of quick keys and telnet type sessions are that they do not look pretty to the eye, and require a day or two more training.  A mouse and windows, even a tablet is useable from day one, so sells, but you can never speed up data entry, and you have to LOOK at the screen.

    UK Primary care doctors and their staff have low turnover, so training is less of an issue for them compared to hospitals, the latter usually employ data clerks.

    The UK National Health service was at first going to scrap all current GP systems, but slowly the realisation is these G.P. systems have working infterfaces, honed in over 20 years.  

    Microsoft has a contract to develop an NHS Interface, but I gather it is based on New Office formats.

    I vote for quick keys and telnet type sessions!  Microsoft could impose the standard of those quick keys in a medical environment.  Then speech recognition could be used within those sessions.

  12. Roger Weeks says:

    most of the ideas here are out of date. The most important factor has got to be the medical language in some recognizable hierarchical form. Here in the UK all primary care physicians are using Read 2 automatic encoding in the Electronic Patient Record – thats 50,000+ physicians across the nation! Some of us are also using Weed Problem orientation. The GP data set which maps to ICD9 and ICD10 also contains symptoms, observation tests iomages and results, procedures, administration (e.g. referals, certification) as well as drugs, therapies, dressings and appliances in addition to diseases. Thus we can create comprehensive medical records electronically just as taught at Medical School – more to say no space

  13. Roger Weeks says:

    Ok here I go The big push to computerize the NHS has totally ignored the GP contribution and is starting again using the untried untested SNOMED CT which has Read 3 (not 2) in it and the hierarchy is not coded but separately stated as patents and children. Forf the input interface I recommend trialling Visual Read type images where the user simply touches the ‘body part’ on the screen and proceeds down a hierarchical set of drop down boxes to find the various concepts needed.

    I personally am working on extension for Read 2 particularly in the area of ‘mind’ mental health and psychology as there is adearth of needed concepts and terms also in SNOMED CT     Roger Weeks

  14. HealthBlog says:

    I’ve called "data input" the final frontier in clinical computing. In fact, data input has been a frequent

  15. I've called "data input" the final frontier in clinical computing. In fact, data input

Comments are closed.

Skip to main content