PIH EMR USER INTERVIEW
Date: 27 Mar 2008 (in person)
Helped guide the development of the PIH Peru EMR 1.0, expert user
What do you use the EMR for?
Data quality, generating reports, exporting data to SAS for analysis
What works well (and why):
- It’s intuitive (though she admits she may be biased)
- Great that it’s web-based & she can access it from anywhere
What doesn’t work well (and why):
- Can’t export data that’s not patient or group of patient-based (such as dst results from a certain time range)
- There are ambiguous status fields that grew organically from the similar, but overlapping needs of different programs.
Going forward, what are important things to consider for OpenMRS 2.0?
- Need to have clear documentation for concepts so researchers can find what they’re looking for and then know what they’re looking at. Perhaps a way to view a key that shows the mapping of a particular form’s fields to its concepts (w/ detailed descriptions)
- Absolutely need to have an easy, flexible way to export data – any format is fine, really
- QC checks would be great, double entry consolidation, audit trails
- Might be useful to have day, month & year separate so that data enterers don’t have to make up dates if they don’t know them.
- Links to records of household members/other contacts
USER TESTING NOTES
Existing cohort builder –
Task 1 – find all adult patients.
skipped over shortcut, went straight to patient attributes, filled a dummy age into upper age limit (didn’t think she could leave it blank) was able to get results quickly
Task 2 – find all patients w/ CD4 count > 200
Wasn’t able to complete. Clicked through a few tabs, looked for something under Encounters. Didn’t understand Concept/Obs so ignored that tab
Comments: generally not very intuitive, confused by drug vs drug set – assumed that drug set would limit her drug choices, didn’t know what “workflow” meant
Mockup #8 Cohort Builder
Comments: Generally more intuitive. Nice that you can put in multiple filters at once. The middle column that shows the number of results for each individual filter is not intuitive. The numbering of those filters to use in the Boolean editing box isn’t obvious.