The Wa State Senate put some creative thinking into Senate Bill 5501 by calling for a private – public partnership for implementing a state-wide HIE.   It calls for a “lead organization to coordinate development of processes, guidelines, and standards” to improve patient access to healthcare information, implement methods for secure exchange of clinical data, and serve as a champion for security and privacy.  The lead organization must be self-sustaining (without state funding) but is encouraged to seek federal funding through the ARRA.   They have named OneHealthPort as this “lead organization”.

Lots of good information here:

http://www.onehealthport.com/HIE/index.php

http://www.onehealthport.com/pdf/5501.pdf

I think we need to look beyond this debate of CPOE vs. BCMA to the implications of “meaningful use” of EHR. Although the definition of this term is still under development, it seems clear that both CPOE and BCMA will both contribute to help close the medication management loop and move hospitals toward the goal of “meaningful use”. Hospitals are being incented to adopt certified EHR systems that incorporate both CPOE and some form of “computer assisted administration”.

Jason Hess of KLAS touts CPOE adoption as key to achieving meaningful use. In fact, CPOE is a significant component in the Meaningful Use Matrix. By 2011, hospitals must use CPOE for at least 10% of orders, and Physicians practices must use CPOE for 100% of their orders.

Although there is much controversy about the efficacy of BCMA, I would argue that the shortcoming is not in barcoding per se, but in poor implementations of first generation BCMA products. RxInformatics talks about how the current generation of BCMA applications did not meet expectations due to “design and implementation faults and resulting staff workarounds that mitigate the efficacy of barcoding”. I would conclude that this doesn’t mean BCMA or barcoding should be abandoned, but that there is a huge opportunity for improvement. The first generation of systems clearly fell short of what was needed.

Prior to July, the Meaningful Use Matrix called for “medication administration using bar coding”. Unfortunately, the ONC has since revised the matrix to call for “closed loop medication management, including eMAR and computer assisted administration” without any reference to barcoding. In my opinion this is a step backwards. The ONC has been influenced by inconclusive studies done on first generation BCMA systems. RxInformatics states that these systems have major design flaws that affect their efficacy in improving patient safety. Meanwhile, they admit that barcoded medication systems reduce Pharmacy dispensing errors. If barcoding works in the Pharmacy, why not at the bedside? I would argue that there is a challenge ahead for BCMA vendors, but that barcoding at the bedside should not be abandoned. As a principle, in order to completely close the medication management loop it seems obvious that medications must be scanned upon administration as well as upon dispense.

Medication non-compliance is the failure of a patient to take the correct drugs on time and in the prescribed dosages. Non-compliance can be life threatening to patients and is extremely costly to both patients and the industry. It is one of the leading causes of death and hospital and nursing home admissions in the nation. For the elderly, Medicare is most often responsible for paying for care; therefore, these costs are directly passed on to the public who fund the system.

The New York Times calls non-compliance the world’s “other drug problem”. (1) Statistics on non-compliance are staggering. An article in NCPIE Prescription Month states that 1/3 of patients take all of their medication, 1/3 takes some, and 1/3 never even fills their prescription. (6) According to the National Council for Patient Information and Education, about 50% of the 2 billion prescriptions filled each year are not taken correctly (7).

The costs of non-compliance are devastating to both patients and payers. Tragically, noncompliance causes 125,000 deaths every year in the United States and leads to 10 to 23 percent of hospital and nursing home admissions (2). An article published in the American Health care Association Journal states that 23% of nursing home admissions are due to noncompliance costing $31.3 billion / 380,000 patients (3). The Schering Report estimates that 10% of hospital admissions are due to noncompliance, costing $15.2 billion / 3.5 million patients (4,5).

1. The New York Times June 2, 1998
2. Smith, D., Compliance Packaging: A Patient Education Tool, American Pharmacy, Vol. NS29, No 2 February 1989
3. Standberg, L.R., Drugs as a Reason for Nursing Home Admissions, American Health care Association Journal, 10,20 (1984).
4. Schering Report IX The Forgetful Patient: The High Cost of Improper Patient Compliance
5. Oregon Department of Human Resources, A study of Long-Term Care in Oregon with Emphasis on the Elderly March 1981.
6. Hayes, R.B.NCPIE Prescription Month, October 1989
7. National Council for Patient Information and Education.
8. Rosalynn Carter Institute of Georgia Southwestern College.

Heard a story I bet you hear all the time. My partners brother in law came down with pancreatitis related to a gall bladder issue. To treat it they had to stabilize all of his systems in an ICU for about 12 days before the pancreas would stop destroying his innards, let him get strong enough so they could do surgery. During his ICU stay, he had a bad (crash cart, paddles, bad) reaction to a pain medication. At this institution ICU charts are computerized and are not at the bed. Next day, the nurse on duty gives him the SAME pain meidcation and again he is on the respirator, crash cart etc. Wife threatens holy war if they dont get it together so it does not happen a third time but he might have died.

The spreadsheet has three tabs, “Values Matrix”, “Cost Matrix”, and “Cost-Values Matrix”.  Begin by naming your features/products that you wish to prioritize across both the x and y axes.  Then select a scale.  I’ve found 1-7 using odd numbers to give the best spread.  1 indicates identity (the two items have the same value) and higher numbers indicate higher importance.  Have your group pairwise rank each item and assign the horizontal item a relative value.  Then give the vertical item the reciprocal.  For example, if your group says Feature1 is 7 times more important than Feature 2, write a 7 in cell C3 and a 1/7 in cell B4.  When you are finished, the spreadsheet should show you the relative value of each priority and this will be graphed in the pie chart.

Now do the same for the cost matrix.  The cost-value matrix will show you the intersection of cost and value.  Generally, you will end up selecting the high value, low cost items, although you may select some high value, high cost items as well.  The results do not dictate yoru final decision, they simply guide your decision to give you the most “bang for the buck”.

Good luck!  Let me know how it goes if you try this out!

For more information, download this article from IEEE: A Cost-Value Approach for Prioritizing Requirements

After all this is through you get to sit down at your desk for awhile, let your frequent flyer accounts rest for a bit, and you begin writing these down as requirements. You end up with literally volumes of ideas, some good, some fair, and some out of the question. It’s pretty easy to extract a set of valid requirements from this superset and eliminate the duds. But then that’s where the hard choices begin. You have limited time, resources, and money so how do you decide which requirements to include in the first release and which to defer? You are scratching your head, because every contributor has said that their idea, problem, or requirment is important.

That’s exactly where a technique called “Cost-Value Analysis” comes into play.

First, try to group your requirements into a limited number of higher level “themes” meaningful to all stakeholders. Try and keep the list down to no more than 10 themes, otherwise the task becomes too tedious and time consuming. Then do a pairwise comparison of both the value of each “theme” versus all the others and the cost of developing each “theme” versus all others.

The audience for the value comparison may be a group of sales people and product specialists or customers. I prefer to use internal folks for this since they have been exposed to many different customer situations and will therefore have more of a global view of the most valuable features. On the other hand, individual customers will focus on what is valuable for them specifically. You may want to use value comparison with customers as a validation step rather than a discovery step.

The audience for the cost comparison should include managers and team leads from the Product Development group including hardware and softwre development, QA, and Technical Writing. Optionally, leadership from other departments may be included as necessary. These folks should be able to provide rough estimates of the cost of developing each product “theme”.

Stay tuned for Part II where I will discuss the actual method of the cost-value analysis and provide examples.