View source for A pragmatic Approach to Implementing Best Practices for Clinical Decision Support Systems in Computerized Provider Order Entry Systems

This is a review of Gross's 2007 article, A Pragmatic Approach to Implementing Best Practices for Clinical Decision Support Systems in Computerized Provider Order Entry Systems. <ref name="gross 2007">A Pragmatic Approach to Implementing Best Practices for Clinical Decision Support Systems in Computerized Provider Order Entry Systems. Peter A. Gross, MD, David W. Bates, MD, MSc J Am Med Inform Assoc. 2007; 14:25–28. DOI 10.1197/jamia.M2173. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2215068/</ref>

== Introduction ==

Four years later, Dr. Bates still needs to stress his Ten Commandments for [[CDS|Clinical Decision Support (CDS)]] (1).  The present article emphasizes the steps between design of a ‘computer program’ and implementing a CDS system.   The connection between the record and the order goes through the physicians hand via his eyes and brain.  The process is called human factors engineering and it is the last function of the three essential endeavors that Gross and Bates discuss to become a mature field. (clinical, technical, human factors). <ref name="gross 2007"></ref>  This article stresses that, like any other computer program, [[EMR|Electronic Health Records (EHR)]] are just another program and Garbage In equals Garbage Out.  The programs depend heavily on clinician’s behavior, and incentives or impedances in the very program and layout affect how the clinician behaves.  

In [[CPOE|Computerized Provider Order Entry (CPOE)]], the accuracy of the CDS drug-allergy checker depends on how fastidious the clinician was on eliciting the allergy history and recording the allergy history.  This in turn depends on whether the allergy list is coded or free text, on how easy it is to find the allergen in the list, how well defined the allergens are and how complicated the list is, and whether the allergy module allows the separation of allergies like airway failure from sensitivities like nausea from codeine and whether the clinician finds it useful enough to spend his time documenting that.  Part of this clinician behavior is design of the program, part of it is that factor called clinical leadership.

As CPOE and CDS becomes more sophisticated, orders for drugs and even labs will be checked against diagnosis.   When does the clinician enter the diagnosis, before or after he writes orders, before he admits the patient or after he does the discharge summary?  The utility of order checkers that check against diagnosis depends on workflow, community practices, and the individual.  If the CDS works against the previous admission or the outpatient record, how well did the admitting doc, the attending doc, or the billing office define the diagnosis?  Did they use free text or ICD9? Did they code to the fourth or fifth digit? Did they just put in a ‘close’ answer to be able to order a lab when the most accurate diagnosis was just too hard to find?

Making the right thing the easy thing to do is the standard to which all clinical programs must be held.  Careful tedious attention to available orders, how many are presented to the ’novice’ and which are reserved for the full list of interventions or tests, and how the orders are represented; affects how likely the clinician is to order the right thing.  Complete ‘order sentences’ with default entry of the most common doses, routes and times can be produced. Markers on the choice list can show which meds are on formulary or which meds the patient is allergic to even before they are chosen.  Hyperlinks can link choices to the formulary or to the Physicians Desk Reference.  Drugs that will interact with the present meds could be flagged or grayed.  Color-coding can be used much more than it is now. 

When designers start to plan CPOE and EHR programs, they must obey rules of human factor engineering.  “Human perception, memory, reliability, and other aspects of cognitive psychology should be carefully considered." <ref name="gross 2007"></ref>   Somehow, designers and programmers accost the senses without providing needed information.  A Review of ‘The Design of Everyday Things” by Donald Norman should be mandatory for the layout people. 

== Implementation ==

Next the authors discuss the ways that implementation itself is important to the function of the EHR.  “In determining the overall success of the project, the process of planning for and executing a CPOE and CDS implementation can be as important as the selection of the system to install." <ref name="gross 2007"></ref> Health care is a complicated system, and inserting a change into one part of the flow creates changes n other parts.

CDS systems create alerts when they are programmed to; for a defined error condition.  It could be a drug allergy interaction, or a drug-drug interaction or any other definable parameters that the programmers and designers have determined should not happen together.   Drug-drug interaction checking systems fire frequently, often for insignificant reasons, like the oral Prednisone--topical Aristocort situation of a patient with severe poison ivy.  The system does not have the necessary programming to separate topical drugs from oral drugs.  Likewise, other minor or insignificant interactions that nonetheless fire the rule and interrupt the clinician, leading to a condition called Chronic Alert Fatigue Syndrome (3). The more often clinicians see alerts that are useless the faster they learn to ignore them.  “This CPOE feature often leads to alerts for so many low-clinical-relevance interactions that clinician-users ignore the most critical interaction alerts due to ‘information overload’ or ‘inability to recognize the needle in the haystack." <ref name="gross 2007"></ref>

CPOE reduces some mistakes of the pen, trading illegibility for possible proximity errors when drugs are chosen from a list.
CDS reduces some mistakes of knowledge or attention by coaching for allergy or drug issues (or pregnancy, physiology, age).
The common field where the computer, orders, care site and patient interact is in the clinician’s brain.  This component must be treated carefully.

== Dr Bate's 10 Commandments ==

The best way to conclude this review is with a short version of Dr Bates Ten Commandments: <ref name="gross 2007"></ref>

# Speed is Everything
# Anticipate needs and deliver in real time
# Fit into the User’s work flow
# Little changes can make big differences—improve usability to ‘do the right thing’.
# Physicians will strongly resist stopping
# Changing direction is easier than stopping
# Simple interventions work best—simplify guidelines
# Ask for additional information only when you really need it
# Monitor impact, get feedback, and respond
# Manage and maintain

== References ==
<references/>

1. Bates DW, Kuperman GJ, Wang S, et al. Ten Commandments for Effective Clinical Decision Support: Making the Practice of Evidence-Based Medicine a Reality. JAMIA 2003; 10:523-30.
3. Greengold, NL. Chronic Alert Fatigue Syndrome: An In-Your-Face Dilemma. P&T 2005; 506-507,511.


[[category: OHSU-SP-07]]
[[category: Reviews]]
[[category: CDS]]