Transforming the Supply Chain landscape with AI and Machine Learning

Transforming the Supply Chain landscape with AI and Machine Learning

Mark Welch, Senior Vice President, Supply Chain, Novant Health

Mark Welch, Senior Vice President, Supply Chain, Novant Health

The healthcare industry is evolving rapidly thanks to years of investment in foundational technologies and processes, particularly related to electronic health records and diagnostic platforms. In many health care systems, back-office functions have had to give up investment to fund the patient-facing solutions, and it can be easy to get complacent and rely on legacy procedures indefinitely. Aware of this risk, Novant Health has been proactive in updating our back-office processes and data capabilities so that our physician partners can make better and informed decisions at a rapid pace. New advancements in computing and the proliferation of data sources have allowed us to boost efficiency in automation. As a result, clinicians, team members and end-users feel empowered to swiftly engage in procuring the required products and services at every level to quickly achieve their desired outcomes. At Novant Health, we had to develop better ways to access healthcare procurement systems and give providers what they need to support patients and continue tracking expenses.

In 2012, we started the Clinical Variation Reduction program with the goal to standardize recognized best procedures through identifying opportunities to decrease clinical variation across Novant Health’s facilities. During this program, we found new ways to look at resource utilization and costs of care. The focus and intent of the program was to provide safe and effective quality care in a cost-effective manner. In gaining enhanced visibility to variations in costs and quality, we created positive change with our vendor community and physician partners. We did this with better data and a manual process to share information in a progressive and transparent way. This old manual process was very time-consuming, but it taught us a lot in terms of where to find the data and how to translate it in a way that allows us to gain trust with all partners. Transparency and data integrity are the cornerstones of our program. With our digital product solution team, we are now on a fast journey of automation with a goal of employing real-time decision making tools that will lead us into advanced automation, and eventually on to automation driven by artificial intelligence (AI).

"When it comes to Novant Health supply chain, we want to be a center of excellence for analytics and predictive behavior"

Facilitate with Influence

Even with cleaner data, the ability to translate with visualization requires the presence of a capable workforce. With our new automation of data collection processes, we can push our team towards AI-based solutions that focus on patient outcomes, revenues and pre-procedure costing models. One area to overcome is the four to six week lag time on the supply expense analytics related to DRG or CPT codes. We use that data to facilitate conversations and influence future behavior within our provider community. Using basic analytics, we can break down the associated raw data and share it with all involved parties. For example, we overlay cost information by procedure and normalize it with our Case Mix index. Or we can overlay of read mission rates or length of stay data in relation to supply expense. These views provide an opportunity for us to facilitate conversations with a visualization tool which shows several data points that can compare outcomes and cost. Our supply chain teams are at the table facilitating the data and our physician partners are driving the clinical discussion and focus. This partnership is strong and results in a sustainable change that creates trust and value while promoting quality and cost reduction.

Now that we have an established methodology to create these visualizations and conversations, we are exploring AI-based solutions toquickly address our needs, as well as develop our ability to predict outcomes and cost before the case begins. This phased approach to AI adoption allows us to gain understanding of potential and wide-spread buy-in from our colleagues.

Easy AI adoption can also be facilitated by introducing simple bots for accounts payable and other back office processes. As our team members feel more comfortable interacting in conversational style with AI-based solutions, their appetite for using more sophisticated solutions in the clinical setting increases.

A New Wave

Collecting data real-time in hospital procedural areas can be a challenge due to the lack of standardization with barcoding technology or adoption of data standards. There are a few technology companies trying to solve this issue. RFID is one pathway and another option is sweep barcode reading technology. Barcode reading is not new, but technology and ease of use is advancing in this area. We are implementing a barcoding scanning technology that can read multiple languages and this system will allow us to capture more data consistently.

With this newly captured data, we can advance our supply chain digital world—and we continue to explore AI-based solutions to improve delivery of care, both in the back-office and in a clinical setting. We continue to invest in information management and analytics talent in an effort to better understand the data we collect and how to translate that data into actionable decisions or processes. With more profound insights into the data outputs received, we can help team members and physicians streamline their daily activities.

Earlier in our journey, our data was not consistent and reliable. Now that we have a strong information management capability and a broad understanding by the team of what’s possible, we are positioned to leverage AI in all facets of our organization. With advanced analytics we will be able to predict things like cost and outcomes prior to a surgical case being scheduled, and with AI we can allow for all of those moving parts to be handled automatically. This will be the game changer. This goal and vision leapfrogs the idea of getting a cost printout at the end of the case. We certainly will look and study data post procedure as a means to confirm our digital strategy of predicting outcomes. This post look back will also enhance our ability to train our AI and help the machine learn to become even better.

Key Take Away

A key take away is to develop your vision and ask your team what they want to accomplish and why. Just putting together a nice colored dashboard is not the answer. It is much more than that. You have to start with a vision for decision-making and in that vision, clearly define what you are solving for. It may be quality issues, cost reduction, volume growth or all the above. Then create a phased approach, in which manual processes give way to automation and eventually to learning algorithms for optimization goals to be met. Regardless of what you are solving, you will want to make sure you are in alignment with your health system’s mission, vision and values. Your digital vision also needs to align with the system-wide digital vision sponsored by the executive team.

This is certainly not an overnight journey, but waiting until someone else solves it is not an option. Action is needed now to establish a program’s foundation so you can build on it and create your robust AI solution set.

Weekly Brief

Read Also

Raising the bar of Accountability for Healthcare Leaders with...

Elizabeth L. Holder, Ed.D. System Director, Patient Experience, SSM Health

Is Chat GPT Code Red the Start of a New Epoch?

Jonathan Witenko, System Director, Virtual Health & Telemedicine, LeeHealth

The Evolving Realm of Pathology

Stephanie Whitehead, Executive Director of Pathology, University Health

Technology Innovations in Hospital Medication Management

Carol Chow, MPharm / RLSGB Head of Pharmacy, Gleneagles Hospital Hong Kong Founding Director of Biomedical Impact Association