Tuesday, February 9, 2010

EHR Usability

In my recent Top 10 EHR Barriers blog, I identified usability of EHR software as a key issue.

It appears that NIST, with ARRA funding and ONC guidance is doing something about it.

This Sources Sought request is to identify organizations with the capabilities to accelerate usability testing and guidance. Specifically:

1. Tasks, Users, and Systems Research Component
a. Perform cognitive task analysis of the health care delivery task environment
b. Determine user characteristics
c. Perform human factors analysis of representative sample of existing HIT systems
d. Develop hypotheses about critical HIT design features
e. Evaluate hypotheses with usability testing

2. Organizational Usability Processes Research Component
a. Study existing HIT industry usability practices
b. Investigate current usability evaluation methods being used in healthcare information technology as well as other areas
c. Conduct a workshop of acknowledged experts in HIT usability and have them analyze the findings of steps 2a and 2b.
d. Using the recommendations of the experts in step 2c and the principles identified in steps 1a through 1e, create a well-defined usability evaluation process that includes a roadmap for implementation and indicates where policy decisions are required.
e. Evaluate a representative sample of HIT systems using the evaluation process developed in step 2d
f. Develop a common format for HIT formative and summative usability test reporting

3. Development Methodology
a. Define, develop, and scope requirements.
b. Define and develop the project approach and description.
c. Adopt and refine methodologies following standard usability engineering principles.
d. Define and manage organizational requirements.
e. Perform cognitive task analyses and develop user profiles.
f. Conduct human factors analysis of systems and do hypothesis testing.
g. Identify existing usability practices in HIT development.
h. Organize and facilitate HIT usability expert workshop.
i. Develop, refine and harmonize usability and certification principles and frameworks.
j. Provide documentation and communication.

There is not yet a solicitation for funding, but I'm confident that funding opportunities will be announced soon.

Usability must be a key aspect of the national EHR program. Clinicians note that current products are hard to use/learn and this lack of usability can lead to errors. Senator Grassley's investigation of the current state of EHR products is in part related to usability issues.

My own experience is that usability is journey. Several years ago when the Joint Commission asked hospitals to implement medication reconciliation for all transitions of care, my development groups built innovative software that leveraged inpatient, outpatient, ED systems, and Surescripts/RxHub medication history to support the process. It took us a year to get the usability right by engaging stakeholders, studying the workflow, and iteratively revising the user interface. Only after extensive usability improvements did the application get deployed and adopted.

I welcome the NIST work and hope that we can develop objective usability metrics as part of the certification process. Software should be capable and usable to qualify for certification.

1 comment:

Robert Schumacher, Ph.D. said...

EHR Usability as Process and Attribute
Dr. Halamka makes some great points. Usability is confusing. It is often described as an attribute – “that application is really usable.” However, it can also be viewed as a process. Let's look at usability as an attribute first.

We can say that one system has higher or lower usability than another. This usability attribute describes the ease with which people can use the system to achieve a goal, and consists of three measurable components: efficiency, effectiveness, and satisfaction. For instance, we can measure how many transactions happen in a given period of time or the number of errors that are made. Ultimately, the results of these measurements are interactions between design factors and properties of the user. Measurement leads to diagnosis; diagnosis leads to interventions that, in turn, we hope, increase the usability of the system.

On the other hand, the “usability process” or how we make an application more usable is, as Dr. Halamka describes, a journey. The journey relies on understanding the science base of human factors and engineering psychology, as well as the research tools from behavioral sciences. Knowing what needs to be changed does not mean we know how to change it to improve the design. For that we need designers skilled in the art and science of user experience design. Patient safety should not be compromised because a junior programmer - given responsibility for user interface design - did not comprehend the serious implications of not showing a full drug name in a drop down box.

The main point here is that in order to get highly usable systems one has to go through a usability process that relies on (1) understanding and respecting users, tasks and environments, (2) setting user performance/satisfaction objectives, (3) designing a user interface that meets those needs, (4) testing how well the tool meets the needs, and (5) iterating the design and test until performance objectives are met.

We conducted a content analysis of RFPs for HIT systems that were publically available on the Internet. Of the 44 RFPs that we examined, 21 had some mention of usability. However, only six of them had anything beyond a cursory mention (e.g., ‘the system must be user friendly’). Only one stood out as having done an outstanding job in specifying usability requirements/criteria of vendors – a dental society: www.nnoha.org. So purchasers either don’t ask or don’t know how to ask how to evaluate usability. Further, we surveyed dozens of EHR vendor Web sites and each vendor made unsubstantiated claims about how easy to use their application was. Lip service is easy.

EHR purchasers and vendors are caught up in the ‘kitchen sink’ model: the more features the better. Without a good usability process, there often is an inverse relationship between the number of features and usability scores. Until recently, certification and rating systems only focused on functionality.

But lately, we have seen a palpable shift: the spotlight is now on usability. AHRQ has sponsored some good research and commentary by recognized experts. NIST and ONC are moving to develop a usability framework. And recently, CCHIT has taken moves to rate usability. To move the needle further, purchasers should move focus from features and functions to usability; they should demand outcome measures from user testing so that they can compare systems. Purchasers should also require that user performance measures be built into service level agreements.

Usability is not a difficult concept. Users want it. Vendors must understand it and embrace usability process models.

Robert M. Schumacher, Ph.D.
Managing Director
User Centric, Inc.