Top 3 tips for implementing artificial intelligence that gets ROI
We’ve all heard stories about how artificial intelligence (AI) can come into your health organization and drive innovation. They may seem promising on the surface, but how can you connect this promise with everyday care management, and what does it take to be successful at implementing the right solutions?
Previously, we’ve discussed some of the areas where AI can make the most impact, including predictive care guidance. What makes predictive care such an attractive place for many to start with AI is that it directly relates to what all health organizations see as a primary goal: the effective and efficient care of patients.
Still, many organizations are hesitant about investing in technology without understanding how to get a return on their investment. Installing new technology can result in a struggle to gain adoption among coordinated care teams, or a disruption of existing care process. Let’s take a look at how you can drive effective AI initiatives by evaluating what initial use cases to pursue, as well as developing innovation champions that drive change and user adoption.
1. Understanding when a business objective is an opportunity to turn to artificial intelligence
Very few healthcare leaders wake up and think “I need artificial intelligence.” Instead, their goals often stem from real problems that must be solved, such as “we need to improve health outcomes,” or “we’re losing a lot of revenue on claim inaccuracies.”
Artificial intelligence is part of a set of transformational technologies, including big data and cloud, that are becoming essential to address strategic problems. The advantage is increasingly clear—according to McKinsey, AI has up to 44% more incremental value over other analytics techniques in healthcare. But you don’t have to jump straight into the deep end to get results. These four primary scenarios in operational and financial analytics are typically a good place to start.
Analytics in the emergency room (Operational analytics)
Because ER visits are costly and a large source for patients in need of timely and critical care, using AI to monitor or predict patient load, required care, and staffing capacity can influence cost savings, care outcomes, and improve KPI achievement.
Predicting length of stay (Operational analytics)
This is typically a large source of cost overruns for hospitals. Using AI, you can get ahead of problems through preventing readmission and complications that extend length of stay, improving patient outcomes and reducing time hospitalized.
Preventing fraud, waste, and abuse (Financial analytics)
An estimated $455 billion is lost every year from global healthcare fraud, waste, and abuse (FWA). AI is proven to be excellent at identifying abnormal behavior, locating the source of wasted funds, and suggesting process improvements to prevent FWA.
Claim denial management (Financial analytics)
Denied claims are a significant challenge for health organizations—it’s estimated that more than 9% of annual claims submitted are initially denied, accounting for more than $262 billion dollars per year. AI can help identify and tag claims likely to be denied, or even expedite claim resolution over time.
By starting with these scenarios, organizations can easily apply AI to existing processes like financial analytics or enhance care team coordination and effectiveness with minimal disruption. Internal champions can also be useful resources for selecting your initial use cases.
2. Assembling your plan for success by setting objectives and building champions
Though the benefits may seem clear, the crush of everyday demands makes it difficult for providers to stay current with today’s rapidly evolving digital landscape. Despite taking steps forward on the road to transformation, almost two-thirds of healthcare business leaders admit they are struggling to keep up with the relentless pace of change.
As a result, it’s critical for transformation initiatives involving AI to be planned effectively so they’re manageable and set up to be successful. Essential parts of the planning process are determining your objectives, evaluating your data and ensuring prospective solutions will integrate seamlessly with your clinical workflows.
Setting attainable objectives starts with getting the right innovation champions in the same room to discuss different perspectives and priorities, and drive buy-in. The ideation process should include people who understand the relationships between cost of care, clinical quality, and technology, such as the CFO, CMIO, CMO and CIO. These champions should ask questions such as “where are we currently experiencing the biggest challenges?” or “where could we make operational or financial improvements using AI?”
Once objectives are in place, you need to think about your data and workflows. A consistent data approach gets better results, but many on-premises systems don’t integrate well with new technology, impacting data quality. Given the myriad ways that data is created, capturing data that is clean, complete, accurate, and formatted correctly for use in multiple systems is an ongoing battle. Forward-thinking organizations are adopting a cloud-first perspective, enabling them to handle increasing data volumes, minimize the impact on teams and enhance existing processes to facilitate intended outcomes.
Ultimately, transformation requires engaging stakeholders early to understand key objectives, evaluate current data maturity and align on how AI could be integrated into existing workflows. Once your strategy is in place, it’s time to focus on change management.
3. Supporting new technology with new behavior through effective change management
Achieving widespread adoption of new technologies is a must for any organization trying to meet transformation objectives, yet it can be an uphill climb. For example, more than 75% percent of surveyed physicians say electronic health records reduce their productivity.
The secret to adoption lies with your change management approach. Doctors, nurses, and clinicians live in an evidence-based world. If they aren’t presented with clear information that demonstrates results, they’re unlikely to support new methods. Involving them early in the decision-making process can aid considerably in earning evidence-based trust, and these same clinicians can later serve as champions during implementation to drive adoption.
Like anything, adoption of new behavior is a matter of incrementally building trust and competence over time. An effective phased strategy to implement AI in your health organization might look something like this:
- First, stage a proof of concept (POC), which will take your data and apply AI in a test environment to give you a good idea of the solution’s capabilities—for instance, whether you can improve over more traditional analytics methods in minimizing length of stay.
- Next, conduct a limited live pilot where doctors and clinical teams provide continuous feedback on the solution and how it connects into their workflow. As the ones who will interact with the system, they’re in a great position to evaluate the user experience, data management overhead, and overall effectiveness.
- Before implementation, share the POC and pilot results with leadership and care teams across the organization to continue building trust and confirm the solution delivers the desired value. Be honest about challenges as well as opportunities. After implementation, continue discussing the solution and potential improvements in weekly syncs or care team huddles, leveraging your champions to influence adoption and identify issues.
With the right stakeholders, the right process, and the right objectives, artificial intelligence can reduce overwork among your providers, improve patient outcomes, and help transform care delivery.
Drive your transformation leadership with AI
This is just a brief overview of how to start approaching AI-driven transformation. What matters most is creating a strategic implementation process that addresses your organization’s unique needs and brings together the champions for both patients and the business. If you’re interested in learning more about how you can get started with AI in your organization, check out our eBook.
