What’s Happened to Clinical Data Capture Over the Last Five Years? – by John Murphy (CEO)

Five years ago HealthTech, the non-profit research and education organization, published a technology forecast report titled “The Future of Clinical Data Entry.”  mTuitive used this report as a guide for our product development direction.  After five years we elected to take a look and see how we did. 

The critical findings of the report are listed below in bold.  Our assessments are listed below in blue, italics.  How does your organization stack up?   Remember, this was five years ago.

1.  Clinical data entry technologies will decrease documentation time, facilitate outcomes measurement and management, and improve the accuracy and standardization of data.

• The market has been slower to develop than projected.  Some innovative organizations bought into the concept in anticipation of the value, but for the most part it is difficult to change clinician behavior unless there are immediate paybacks.  Early adopters are beginning to buy in but the adoption is still in early stage.

2.  Patients will experience second-order effects of clinical data entry technologies from faster turnover of information and reduced clerical errors.

• We think most patients are still waiting for the demonstrable benefits. In our own experience, surgeons and patients generally still wait days to receive a pathology report.

3.  Job roles will begin to change in the next two to five years as medical editors replace transcriptionists, coders become patient data specialists, and clinicians do the majority of their own entry.

• This trend is only beginning to evolve but is definitely accelerating.  Hospitals are aggressively seeking to reduce costs through the elimination of redundant processes such as transcription and data abstraction.  Coders can only code what the clinician puts in the report so the emphasis is shifting improved documentation.  Coding has become even more important as payers improve their review systems, but at its core, coding is another redundant process that will be made transparent through direct clinical data capture.  The best coders cannot code missing data.  Innovative organizations are transitioning to direct clinical data entry.  The real challenge is to make the process at least as efficient for the clinician as the current alternatives.

4.  New entry systems will facilitate the evolution of comprehensive Electronic Medical Records.

• To our chagrin, we embraced the Tablet PC as the “killer application” for clinical data entry.  Unfortunately, users did not share our enthusiasm.  For the most part, our clinician customers who do their own data entry use a keyboard and mouse.  We have integrated speech recognition into our gross pathology product to accommodate the environmental and human factors issues associated with grossing.  Our approach accommodates the capture of structured data which we believe provides advantages in both accuracy and value of data.  We will extend speech capability into other applications over the coming months.  We believe the greatest enhancement will come in the form of touch screen (kiosk like) data entry where the clinician can tap answers and review results immediately.  Preliminary feedback from our customers has been extremely positive in both surgical and pathology reporting.

5.  IT support for training and system maintenance will be critical to successful technology implementation.

• We could not agree more.  Our most effective and satisfied customers are the ones who invest heavily in training and are supported by capable IT organizations.

6.  Speech recognition technology is likely to carry the biggest impacts for healthcare as accuracy is expected to reach 98% in the next ten years. Diffusion of this technology will depend on advancements in post-processing of natural language (e.g., transcribing “BP 120/80 instead of “b p one twenty over eighty”).

• Speech recognition accuracy has improved and the success of companies like Nuance in the healthcare arena supports the validity of this prediction.  Philosophically, our company appreciates the efficiencies of speech recognition but also recognizes the weaknesses of unstructured dictation.  As a company, we take a modified approach that uses speech to capture structured responses as discrete data.  We have also used post-processing of free-spoken, restricted domain, utterances to capture discrete measurement data for our speech recognition enabled products.

7.  Over the next five years, pen and paper systems will gradually decline in use. The digital pen will become more flexible and act to complement the mouse and keyboard. Virtual keyboards on PDAs will facilitate entry for low volume information such as form filling and prescription entry, and bar coding will reduce data entry time and increase the accuracy of drug and materials administration.

• We have not seen much of the digital pen and do not anticipate great acceptance.   Currently our products are used for high-volume data-entry and are not intended for PDA deployment; however, the “iPhone” phenomena may provide a solution, so we are paying close attention.  Overall the reliance on paper is only slightly less than five years ago but the decline in use of paper-based systems appears to be accelerating.

8.  Nearly all data entry technologies will evolve to exploit eXtensible Markup Language (XML) in the next five years, making data web-compatible and enhancing interoperability of standards and modes of entry.

• Although the trend is underway, it cannot happen fast enough.  HL7 now accommodates XML but the elusive “holy grail” of interoperability, one, single, structured and standard format will not happen.  Healthcare has not learned from the experience of other industries.  While the academics and bureaucrats devote time and other resources meeting, debating and formulating the new standard, the market will make the determination.  It may be too late for many vendors whose desire to maintain a competitive edge has created the opportunity for some unforeseen, but extremely intimidating competitors.  What standards are Microsoft and Google following?  At inception, mTuitive recognized that our applications required interoperability and embraced XML.

9.  In the next five years and beyond, the use of multimodal entry systems will bring average dictation time to diagnosis time.

• In pathology, we refer to this as the single step sign out, where the report is issued at examination with no need for transcription or editing.  In pathology, this is still relatively uncommon.  Speech recognition brings this objective closer to reality and has had the greatest impact, but from the mTuitive perspective, the value of structured data is lost. It is our belief that a flexible, tailored user interface will achieve this objective.  The combination of touch screen and speech will be the optimal solution in most cases. 

Curiously, the report missed perhaps the biggest development, browser (web) based data entry.  The ability to enter data from virtually any computer opens a realm of possibilities for solutions and business models. As a Microsoft development partner, mTuitive had a relatively easy transition to a browser based interface.  The core applications deploy easily under MS Silverlight framework.  This expands options for user preference and also opens alternative business models.