Intelligent Rx

An analytics platform for a pharmacy drug-safety alerting service

Project Synopsis

An analytics platform to help the pharmacy team at a health insurance company better understand the performance of a drug-safety alerting service

The Problem

The pharmacy team at a Fortune 50 health insurance company was using an alerting service to protect members from prescriptions that could be harmful.

When a member visited a pharmacy to pick up a prescription, the service would alert the pharmacist to any conflicts.

If rules were overzealous, they could be producing negative customer experiences (like turning a patient away that could safely benefit from the medication). They needed a way to measure the impact of rules on their 14,000,000 members.

Goals

• An internal, web-based tool allowing the team to better understand the outcomes of the drug-safety rules
• To better understand impact of alerts on members
• Reduce outdated sharing methods of static documents and spreadsheets
• A single “source of truth” to share results with other teams and external auditing groups

Process & Contributions

• Conducted product discovery discussions with pharmacy team and Product Owner
• Defined functionality goals, roles and personas to guide user stories
• Created user story/workflow maps highlighting rule authoring, intended outcomes and areas needing illumination through reporting
• Created wireframes of initial features and workflow
• Created interactive prototypes in HTML/CSS - Code was integrated directly into Angular templates
• Conducted usability-testing sessions

Low-fidelity, interactive wireframes were created throughout project to guide team discussions and conduct user feedback sessions

Outcomes

• A simple, responsive web-based analytics platform to help the pharmacy team understand impact of rules and improve member experience at the pharmacy
• Eliminated long delays in results collection from pharmacy benefits management service - drastically reducing time spent on information collection for auditing groups and other teams
• Reduced network and shared drive clutter and need for historical document archiving
• Tool is trusted by outside auditors as a source of truth regarding how drug-safety rules are performing and protecting the member population
• Project was delivered within budget and on-time. Part of this was due to rapid exploration and iteration processes (focusing on MVP)
• The pharmacy team is delighted with the tool and is eagerly discussing features of phase two with Product Owner

A view of pharmacy transactions with summary information

Transaction detail displays claim, member, physician, pharmacy and drug information. Alerts associated with claim are displayed with relevant information for analysis and auditing

Responsive design was utilized to accommodate tablet users

The accompanying rules authoring system presented an easy way to find and edit rules

Expanded rule in edit mode in the authoring system

We had some fun with the project and I got to exercise my illustration muscles

What I Learned

• Using layout code as our prototyping process reduced design asset to code translation time and energy. It also moved a portion of development time to the design process. In retrospect, we could have used more prototyping tools like Sketch for initial testing and coded product templates later in the project.
• Collaboration with dev team doing sample queries based on user story maps allowed us to design components and information displays more accurately. There were no surprises during implementation phases.
• Simpler language around drug concepts were rejected by the pharmacy services team. During initial user testing phases, I realized they used very data-specific language as a business unit. Product language evolved to reflect their technical business culture.
• I learned a lot about insurance practices around prescription fulfillment and probably more than I wanted to know about First Databank (FDB) Drug data :)
• It was rewarding to work on a project that ensured the safety of 14,000,000 people!

Next Case Study →