(Uploading my personal documentation of "how crowdmed works" soon!)

I started a new development cycle since I started leading UX. I studied the platform thoroughly and carried a small scale user research to understand UX issues on the platforms. Based on the insights collected, I laid out strategy for redesigning, i.e. set up an order in which all the pages/features on the site should be redesigned.

The case page is the most frequently used page and it had the most critical UX issues. Here is a brief documentation of UX issues on this page:

UX Issues with Case Page by Aashka M Shah on Scribd

I fixed these UX issues and also introduced a new way of allocating points (recommending a suggestion) along with point allocation history feature.


Tweaking Point Allocation:

Point allocation (recommending suggetions) is only allowed once a month on any case with a maximum limit set based on detective's rating. This is not allowing detectives to recommend a suggestion at a right time. This is generating either delayed or premature recommendations on the webiste. In the new design, I am allowing detectives to allocate points any time with a maximum limit each month on each case. This functionality tweak will not just enable timely recommendations but make the interface way more simpler for detectives. Also, medical detectives participate on an average 20 cases at any given time. It is really hard for them to remember recommendations they have made on each case, so I introduced point allocation history feature on this page.

We also observed redundancy in the suggestions. Sometimes a specific disease under a spectrum and the spectrum are added as seperate suggestions. To avoid this, I introduced nesting in suggestions. In the new design, detectives can add/recommend a nested diagnosis as shown below.

The most important change is: displaying discussions under suggestion rather than showing under a seprate tab as shown below.

Also, tweaked display card with the list of suggestions on the patient side to match up with new design for case page.

Adding/Recommending/Flagging Suggestion:

Due to limited resources, CrowdMed was mainly focusing on developing web platform. I saw equal traffic coming from mobile and web on Google analytics and introduced mobile first approach. Here is the mobile version of the new case page design:

This is a very important process for the platform as patients provide information about correct diagnosis and most helpful detectives which CrowdMed uses to distribute reward. I sketched out userflows for this process and some of the flows were missing in the design. Also, CrowdMed was not showing how the reward is being distributed clearly to patients. This led to dissatisfaction/confusion in the patients especially the ones who did not get helpful suggestions on the platform.

Shifting success metrics by rewarding most helpful diagnosis instead of correct diagnosis:

CrowdMed used to ask if patient received "probabily correct diagnosis" on the CrowdMed. When patient received correct diagnosis, it rewarded medical experts/detectives who participated in posting and recommending the correct diagnosis. This fuctionality did not allow to reward medical detectives who posted or recommended on helpful diagnoses on the case which did get the correct diagnosis.

New Design:

Now CrowdMed asks for the most helpful diagnosis and rewards medical detectives who participated in it. I had also sketched out solution to reward multiple helpful diagnoses, but we decided to implement it later. Here is the new design for case solving process:

From user research, I found out that medical detectives used to chat with each other through review feature. Crowdmed did not allow two medical detectives to have private conversation to maintain transparency on the platform. This made it really hard for medical detectives to take help from each other and learn how crowdmed functions and medical detectives started using review feature to chat. I re-designed medical detective bio page with messaging features along with review as shown below to provide them a seperate channel for private conversations.