A $1B UX Success
Pyxis (Now Beckton Dickinson) was developing a narcotics kiosk for hospital nurses and demanded an interface that eliminated medical errors and controlled Schedule IV narcotics against misuse. Moreover, the interface needed to foster a user experience brand that nurses wanted to use.
When interviewing for a job, nurses will ask if the hospital uses Pyxis. Some nurses won’t take the job without a Pyxis.
User Research Methods
A quick Heuristic Review identified a number of critical issues with the original design. This prompted a more thorough user observation and task analysis that identified more user-oriented design approaches.
This particular project involved both generative and evaluative methods, such as Heuristic Evaluation, Interviews, Competitive Analysis, Usability Testing and Observations. This effort involved a couple dozen internal resources and over a hundred users.
The competitive landscape was composed of several mediocre solutions to the narcotics control problem and the Pyxis team needed to eliminate the competition.
Target User (Persona)
Nurses on inpatient hospital floors who were responsible for administering controlled substances (narcotics and Schedule IV medications.
Nurses had many patients to attend to and it was difficult for them to remember all of the different medication orders
Patient prescriptions called for specific meds at specific times. A nurse had to administer the right meds at the right time, which could be very confusing.
No Medication Errors!
Medication errors consists of various mistakes including administering the wrong med, administering the right med at the wrong time, giving the wrong dose (too little or too much), missing a med altogether, or mixing incompatible meds (drug interactions).
Nurses are practiced and familiar with many common medications, but this broad knowledge requirement also increases the opportunity for mistakes.
Besides administering medications, nurses were also required to keep track of the medications in the kiosk
UX Design Strategy
The user research highlighted that the nurses think differently than we would expect them to. For instance, when assigning nurses to patients, the initial approach was to select a nurse, then assign patients to each nurse until all patients had been assigned. This was seen as a one-to-many problem. Our observations indicated, though, that an overriding focus is to reassign a patient to any nurse who had recently cared for that patient.
By starting with the patient and then determining which nurse cared for that patient last, the problem changed from a one-to-many to a one-to-one problem
This focus on the patient first approach was referred to as “continuity of care” which became one of the supporting design strategies of the overarching design strategy “No Medication Errors.”
The original design was based on the developers’ understanding of the system rather than empathizing with the users.
For instance, one task required the Charge Nurse to assign nurses and patients. The developers determined that since each nurse had many patients and patients were assigned to only one nurse that this was a simple one-to-many problem. Their solution required the charge nurse to select a nurse and then assign many patients to that nurse.
Continuity of Care:
The observations determined that that solutions was actually backwards to the way nurses thought of the problem. A prevailing focus in assigning nurses and patients I referred to as ‘continuity of care,’ meaning that Charge Nurses try to assign the same nurse to the patients as much as possible. So, rather than selecting the nurse and assigning patients, they typically select the patient and then assign a nurse based on this ‘continuity of care’ approach.
The key to the success of this medical device was removing as much of the cognitive and memory components from the nurse’s work flow as possible. The kiosk is less of a simple kiosk and more of a medication assistant. Keeping track of meds given (or not), alerting when something is amiss, and preventing when the nurse attempts to commit an error.
The Pyxis Medstation is a basic consideration for nurses when applying for a job at a hospital. Nurses are hesitant to accept an offer from a hospital that does NOT have Pyxis Medstation installed throughout their facility. The Pyxis brand is the standard in medical device user experience design to which all others are compared.
Numerous independent studies have proven the effectiveness of the kiosk to reduce medication errors by 90%, almost completely eliminating them. The errors that do occur are reported as human error (meds loaded into the device incorrectly, incorrect dosage by nurses, etc.) with none attributed to the device or usage of the device.
As a side note, Pyxis was sold to Cardinal Health for almost $900 Million within 3 years of the new user experience/interface design.