An important aspect of the DNA of any software engineering organization is agility and the ability to learn, innovate and fail fast. It's even more important when your clinical software solutions focus on vital healthcare areas such as the OR and the Emergency Department.
Build/Hire/Buy vs. Agility in the World of Software Engineering
In a number of cases this agility involves a build/buy/hire decision. With the advent of open source software, especially for startups, there is a greater incentive to build/hire rather than buy. Adopting a paid platform, although beneficial in the short run, typically enforces rigidity in the way the software can be configured and used.
If the engineering organization ends with the build/hire decision, there is a learning curve involved no matter which fork of the road is chosen. Hiring for special skills is extremely important and will add to the technical depth of the team. However, hiring is not as easy as it sounds.
AI Engine to Predict Surgery Case Length
A few years back, TAGNOS wanted to build a AI engine to predict case length for surgeries.
Although several members of the software team had used AI and machine learning in the past, much of our experience focused on recommendation engines. No surprise as that was the domain we were involved in when we applied AI to solve those specific problems. We were working on recommendation engines for learning systems to recommend videos and learning material to users based on various algorithms.
Before we could start hiring, we needed to do our homework. We needed to figure out which sets of tools and algorithms would be a good fit for the kinds of problems we would take up at TAGNOS. Doing this homework would help identify what skills the correct candidates needed and how to intelligently explore their past experiences.
Let us talk about another project before we continue the AI story.
Electronic Health Record Vendors and HL7
TAGNOS interfaces with electronic health record vendors using a communication methodology called HL7.
At the time of the AI Engine example, TAGNOS had a custom HL7 engine when several HL7 mapping tool companies approached us to look at their technology.
We wanted to do a quick proof of concept project to evaluate how the vendors would help us.
Rather than a quick buy decision, this became a quick build project on the TAGNOS side along with an educational component. During the process, we realized that these vendors actually applied archaic coding practices that would require maintaining different code bases for various HL7 versions.
We completed the proof of concept project in a couple of weeks. It resulted in spectacular failure! The code built out for the proof of concept not only didn't work reliably, but we also realized that the basic premises on which the integration was based on was faulty.
Back to the AI Engine with a Proof of Concept Approach
Coming back to the AI discussion, we did a quick proof of concept project in a similar vein.
We ended up evaluating several open source solutions we could tailor for our needs. During the course of the exercise, we obtained test data from a few of our healthcare system customers.
In this case, the result of the proof of concept was so encouraging that we decided to build on our own and, instead of hiring, we sponsored a few of our own software engineers for advanced training.
As you can see from this example, being able to quickly test and prove out technologies is core to any software engineering organization. It's even more beneficial given our focus on healthcare solutions.
Mobile Applications Technology
One more example for you relating to mobile applications technology. Although the ending wasn't as desirable as with the AI Engine project, it has still been useful in the long run.
One TAGNOS team member picked up a new technology called PWA that could be used to create mobile applications. Close to the hybrid applications already in use at TAGNOS, it seemed even better since it required no actual app to deploy from the app store or play store. It was just a mobile web page that would be created as a short cut. PWA seemed friendly from a deployment point of view and we assigned an engineering architect to work on the app for a few weeks.
The initial results were encouraging in Android tests. However, we soon realized that PUSH notifications, a key component of the TAGNOS application, would not work as expected in IOS. Although Android worked working fine, testing on IOS revealed issues.
We decided to scrap the project and chart it as a learning experience.
What Agility Has Taught TAGNOS: Fail Fast, Innovate and Learn
The key to this kind of agility is the ability to have a small group work on these proofs of concepts and be willing to pay the price of failure. As long as they fail quickly, the investment that is needed is relatively low.
Apart from weaning out bad technology, this kind of a process instills a spirit of discovery and innovation in the group and that is more valuable than anything else for a startup.
What has this kind of agility taught your organization?
Thanks for reading.
TAGNOS is the future of clinical automation software solutions with Artificial Intelligence. It is the only platform offering predictive analytics utilizing machine learning and RTLS. This groundbreaking platform leverages historical patient data continuously and adjusts operational intelligence to provide sustainable improvement to both the patient experience and metrics.
TAGNOS provides clinical systems integration, customizable reporting, dashboards, alerts, critical communication with staff and family to improve turnaround times. TAGNOS supports patient flow, workflow orchestration, and asset management.
In the course of 13 months, hospitals see a 12.7% reduction in its overall cycle time - saving an average of 40 minutes from each case and over $1.6M per year - more than 11x the typical investment.