HOW CAN WE? Create a digital solution for something usually done manually and go from weeks to minutes!

---

What was the problem we solved for our user (Lab Directors, Pathologists and the Tumor Board)?

Up until Watson for Genomics was created, getting a tumors genes sequenced, putting them into a database and coming up with clinical trials and drugs for that specific cancer was done manually, taking weeks to come up with the solution for the Pathologist and the Tumor Board, let alone the lab Directors.

Watson for Genomics hypothesized that someone when biopsied a tumor you could take the genomic sequence, and with the help of machine learning and AI that instead of going through thousands of clinical trials and all the drugs, that it could be analyzed and done in 2-5 minutes. Including a filtered list of FDA approved drugs, and drugs that would work for that particular tumor. This would give the Pathalogist, Oncologist and the Tumor board the absolute best information to help a late state cancer patient. Once the Pathologist takes a look and decides which ones to go with, a PDF report is created and sent to the Oncologist in hopes of finding something new for a patient in dire need.

(Below: The application as it was given to my team and myself. Pretty informative, huh?)

The problem to be solved in this case was initially thought to be for the Oncologist, but in reality it was for a Lab Director and Pathologist. The Oncologist, and subsequently the cancer patient was really secondary, and you had to throw in the Tumor Board because they were also a group that consumed the report from (from here on out Watson for Genomics will go by WFG) WFG.

Since most of the companies doing this were doing it manually, it would be a pretty big deal if we could pull this off. Taking something that usually takes weeks to then a few minutes is a UX'ers dream.

How such a big hairy problem was solved

We were initially asked to make what they had pretty. Of course I said we would, but did they mind if we also presented them with a better solution?

1. No one on the team knew anything about genomics, so we started doing the research
2. We did some card sorting.
3. We initially were told Oncologists were the target users.
4. After many, many interviews and workshops we concluded they were the last in line and that lab directors and Pathologists were the real target with the report being what the Tumor Board would eventually see to decide which option was best.
5. We had the challenge of it being in an iFrame and 2 different dev teams in different countries working on it.
6. 3 of the designers on the team were fresh out of Industrial Design school and had no UX or healthcare background.

The real issue as I look back on it today was the PDF. If I regret anything it was not pushing harder to make it HTML and interactive like the application itself instead of it just spitting out a PDF.

When we presented out solution, the PM and team LOVED it. We had made their initial deadline for the "pretty" version, and then we were given the time to do things right. The full blown DUX process and to make things even better we had full access to our subject matter experts!

We held Design thinking workshops along the way to come up with Hills (an IBM DT term kind of like a Needs statement but making sure things are maasurable and time-bound) which stated the Who, What and WOW.

My team worked hand-in-hand with our own Bioinformaticians and Cancer biologists and we did a fair amount of interviewing of cancer patients, Oncologists and Pathologists. We were lucky to also have our subject matter experts within IBM who were also these things we could talk to all the time along with the Sales people they hired were all Oncologists and Pathologists. IBM Design believes in having sponsor users, which allows us to do usability testing easily on a regular basis.

Of course, best laid plans...it wasn't long after we really started that I had to pull everyone else off the project and onto other projects. Watson for Genomics wasn't launched at that time, and for a bit I tried to hand it off to another team. The product lead wasn't happy with their approach and insisted (to my boss) I be put back on, as things had got off track. So I ended up having to truly design the entire product, and had a great time doing so.

How did we measure success?

Once I came back onto the project, I worked super close with out Sales team (who were all current or former Oncologists and Pathologists) and the SME's I had about 8-feet from me. We worked hard at finishing this up to a point we were able to launch the Product in April of 2018. It is always exciting to see something get released to the General Audience intended, but now we really had more people to get real feedback from.

While Watson Health is no longer Watson Health and the product no longer exists we did have our fair share of successes. It was on 60-Minutes and even saved lives.