Case Study
User data and machine learning leads to faster, more accurate prescriptions
UpToDate has a wide array of articles that teach clinicians best practices for both medical topics as well as drugs.
While conducting user interviews to learn more about how clinicians use their mobile devices on the job, the team heard consistent feedback pertaining to UpToDate’s drug content.
Our search team and UX researcher looked further into usage data to help support this plan and to help define how we should approach potential designs.
The data showed that:
A high percentage of UpToDate traffic goes to drug content
A high percentage of all drug searches end in a drug monograph page
A high percentage of drug monograph traffic ends in one of three sections (two dosing sections, one adverse reaction section)
Based on these findings, we hypothesized that with more advanced auto-suggest and type ahead tools in the platform, we could decrease the time it takes to bring users to the relevant information they need.
Provide users direct links to drug monographs, and popular sections at the point of their search.
Support all common user needs, including searching for brand name drugs and drugs with multiple routes (routes are different ways to take a drug).
UX Designer (me)
UX Researcher
Editorial Staff (practicing clinicians)
Product Manager
My role: Take the user feedback, data, and project priorities to lead design to find a solution that easily gets users to drug information faster.
The previous user flow required users to:
Type into a search bar
Search for their typed string [or] click a search suggestion from the search bar dropdown
Click drug monograph link in search results
Locate and click dosing link from outline
Read dosing information
By adding both links to the overall drug page and the specific sections, the user now has a shortened user flow:
Type into a search bar
Click a “Go Directly To” dosing link from the search bar dropdown
Read dosing information
Allowing users to search for brand name drugs and routed drugs (multiple way to take a drug) were also priorities for this project. Given the simplistic design approach, we were able to support these use cases.
Many users search for the brand name of a drug, and not it's generic name. We needed to support both cases, while linking to the singular page that covers both.
Many users search for the brand name of a drug, and not it's generic name. We needed to support both cases, while linking to the singular page that covers both.
Some drugs can be administered in a number of different ways: orally, via an IV, topically, etc. Each may have different rules around how to prescribe the drug, and different doses based on the route. It's critical that clinicians can get to the right dose for the right route quickly, accurately and easily.
The feature was released to a small group of beta users in Fall 2021. Based on feedback collected during the beta testing, slight changes were made and the feature was released to all users in Winter 2021. I had left the company right after the beta release launched.