Plans in Place for the AI Revolution?
There is a lot of talk about AI in our industry – much of it fast and loose. It all warrants a healthy dose of caution, if not skepticism. This is particularly true when listening to anyone who professes to be able to predict with precision what our AI/Machine Learning future holds. Separating fact from fiction, practicality from abstract, requires tenacity.
We know that it will revolutionize many industries, including ours. Exactly how it will play out is a big unknown. One thing you can count on is this: You will need a plan for AI and data management.
Put an effective plan together – ideally with an outside consultant.
A Plan, A Vision
AI and ML are already revolutionizing the healthcare sector, with applications in diagnostics, personalized medicine, patient monitoring, and more. Machine learning algorithms can analyze vast amounts of patient data to detect patterns and predict health outcomes, while in the lab AI-driven tools assist in clinical decision-making and streamline administrative tasks like coding and billing.
But in studying the issue for our industry, another adage comes into play that I wish more would heed: Look before you leap. The competitive pressure is on, but this is a situation you do not want to rush into. Thus, the planning stage is critical. And that starts with questions.
- What are you trying to accomplish/what is the goal?
- What are the tasks to be automated?
- What are the actionable items?
These conversations can be started internally, but including an outside eye will provide a valuable perspective that can lead to considerations, ideas, and solutions that might not be uncovered otherwise. Interoperability and how it all ties together deserves thoughtful analysis. And it should start with the reality that each organization deserves to have its own model that is tailored to fit its needs. This is not a one-size-fits-all situation.
Fuel for the AI Engine
The challenge is to integrate and automate the data with the AI engine. But before we get there, we need to look at your data management. U.S. HealthTek’s founder, Bryan Firestone, likes to use the phrase “slice and dice the data.” That’s what needs to happen, and the data needs to be sliced and diced well for AI to run effectively.
The way to think about data management to feed into AI is the old adage, “garbage in, garbage out.” In the computer science world, it means that flawed and poor-quality information results in defective and poor-quality outcomes. A simple way to understand this is thinking about what is being fed into the AI engine. The data should be considered fuel for that engine, and you want the cleanest, best fuel you can create. An experienced consultant – making sure the data is solid and of high quality – is the foundation of successful AI implementation.
So, while the preliminary step is growing your data “bricks,” here’s a shocker: Many organizations don’t own their data. Not owning and controlling the data is something that is more prevalent than you think. This also leads to an inability to share the data within the lab, and having systems that don’t talk to each other.
You need a partner who knows AI and knows our industry.
Reality Bytes
This is not the time to be esoteric or philosophical – let the kids in the college classrooms deal with the theory of it all. You need practical, actionable items closely aligned with precise goals that will make your lab more efficient and profitable. Getting in the way of that are two scenarios happening in our medical/health care world that warrant caution:
- Vendors experienced in our industry but not in AI; and
- Vendors experienced in AI but not in our industry.
You want that “Venn diagram” sweet spot of getting a partner who knows both, because you want to get this right the first time. Don’t waste resources and money working with someone when there is a chance they won’t deliver the first time.
We’ve been helping labs with the first stepping stone, and that’s organization and growing their “data bricks.” And just recently we talked with a current client who we’re helping with their LIS/LIMS solution, and the conversation turned to what they are doing with their “old” data. They were going to place it in the proverbial trash bin, but I pointed out how valuable it is to hang on to. This turned into actionable items that contributed to further improving their lab, and it’s exactly the mind-frame needed when implementing AI.
AI Hesitations
While we’re seeing some moving too fast and/or in the wrong direction, there’s another group in our industry that has the opposite problem: Be it fear, be it budget, be it a lack of information, they are immobile even when they know those who are hesitant will be left behind. What is keeping some from moving forward?
- Not sure where to start. This is why a qualified outside consultant is key, someone who can assess the situation, build a plan, and help with data management.
- Getting lost in the fog of AI. It’s a new buzz word, and there’s a lot of hype and frankly, misinformation about what it is and what it can do. Take a breath, step back, and do a deep dive (best results come from going through this with an expert).
- Lack of resources. We’re so often used to turning to our own people, but I’m seeing that internal folks in this industry tend to be swamped with the day-to-day, and on something this big, it’s hard to even get your head around it, let alone your hands.
Let’s Learn What Works – And What Doesn’t
Lynn Brock of Sagis will be part of our panel discussion at EWC.
At last year’s Executive War College gathering, U.S. HealthTek hosted a panel discussion on AI featuring Andy Moye and Greg Sorensen. There we got into some specifics of using AI to improve performance in the pathology space. This year, we are again hosting a critical AI panel discussion to go even deeper into providing insight on how to use this technology effectively. It’s exciting because Bryan and I have been recently speaking with Lynn Brock, Chief Innovation Officer of Sagis Dx. Lynn is on the forefront of AI implementation and has a handle on using AI in ways to improve his performance and be more effective and efficient overall. We’ll all learn a great deal from him being part of our panel. Moderating that discussion will be our own Laurie Huard, who has over 35 years in the health and specialty lab industries.
Next month we will preview what that discussion will entail, but if you’re going to EWC in New Orleans (and we hope you are – in fact, we’ll help you get there), plan on joining us at 8:30 am Tuesday morning (April 29) in room Foster 2. This revolution can be overwhelming, but we will make it less so and show you how to take advantage of AI safely and effectively.