Will AI hinder digital transformation in healthcare?
Healthcare digitalization requires integrating new technologies and processes to meet patients' evolving needs, but the process is rife with challenges.
Improving healthcare delivery remains a major priority for hospitals and health systems as issues like chronic disease management and health equity continue to present challenges for providers. Significant efforts to enhance care revolve around gathering and effectively utilizing the data points generated across patients' journeys.
While these tools play a vital role in helping organizations navigate the healthcare data cycle, some have raised concerns that such technologies could also throw a wrench in health systems' digitalization initiatives.
Sowmya Viswanathan, MD, MBA, MHCM, FACP, chief physician executive at BayCare, recently discussed how tools like AI might actually hinder digital transformation -- and what steps the health system is taking to overcome these roadblocks -- in an interview with Healthtech Analytics.
The digital transformation landscape
The healthcare digitalization landscape is unique in many ways, but Viswanathan emphasized that the industry's slow adoption of new technologies is a critical aspect of the conversation around digital transformation.
However, this shifted in the wake of the COVID-19 pandemic, as remote patient monitoring and telehealth became necessary to prevent care interruptions. This shift, Viswanathan noted, was key in helping healthcare providers see the value of new technologies for streamlining administrative work and automating routine tasks.
"When we try to look at 'What are the administrative workflows that we can streamline using AI?' or '[How can we] personalize patient care and create advanced algorithms?' it really started to sink in," she explained. "Then, when we said, 'Okay, we are going to use AI to improve interoperability by integrating it into the various health systems and data sources, and then create that data exchange that is needed for providing appropriate and optimal patient care,' physicians and nurses started to buy into it."
The potential for tools like AI to improve productivity and reduce burnout by serving in an assistive role for clinicians is one of the positive impacts of digital transformation efforts, Viswanathan continued. Despite the hype, healthcare organizations looking to adopt AI to bolster digital transformation and interoperability face various challenges.
One of the biggest challenges is that these efforts are data-driven, which creates data quality pitfalls such as the garbage in, garbage out (GIGO) problem. GIGO is used to demonstrate that the quality of a data output is determined by the quality of a model's data inputs.
In healthcare, this means that using low-quality clinical data in AI and other technologies will result in low-quality -- and, therefore, unusable -- model outputs.
We are big believers that AI technology should be evaluated on an ongoing basis -- it's intended to ultimately reduce healthcare costs and contribute to better health outcomes for patients and communities in general.
Sowmya ViswanathanMD, MBA, MHCM, FACP, chief physician executive, BayCare
"Whatever data is being input into our AI platform is what it will churn out for us. So, if we don't have good data sets to begin with, we don't have high-quality reports," Viswanathan stated.
The second major hurdle is related to costs, as newer technologies typically come with higher price tags.
"AI is still new to all of us, and anytime you see platforms come forward, they bring forward the fact that they [use AI]," she said. "You put 'AI' on the project, and the cost goes up tenfold."
Overcoming this requires interested stakeholders to rigorously assess the capabilities and technical infrastructure built into a given AI platform to determine whether investing in the tool is feasible.
Another significant hurdle to digital transformation is achieving buy-in from care teams.
A key aspect of gaining buy-in within any enterprise is overcoming organizational resistance to change. Viswanathan underscored that tackling such resistance means effectively considering the organization's culture and demonstrating that a tool will alleviate, rather than exacerbate, existing workflow issues through a pilot project.
The ultimate goal of such a project would be to enhance patient and clinician satisfaction, she indicated.
"If it's going to add to the burden of provision of care, then it will be a failure. So, [stakeholders] need to make sure that doesn't happen."
Further, Viswanathan highlighted that healthcare AI is a rapidly evolving field, with new tools being designed and deployed often. Until the market matures, concerns around patient safety, diagnostic errors, data privacy, security and the regulatory landscape will remain difficult to address.
BayCare's AI-driven transformation approach
Despite these challenges, BayCare embraces AI and digital transformation to improve patient care and bolster operational efficiency.
"We are big believers that AI technology should be evaluated on an ongoing basis -- it's intended to ultimately reduce healthcare costs and contribute to better health outcomes for patients and communities in general," Viswanathan explained.
However, she emphasized that the health system's technology investments are made with the knowledge that these tools cannot replace humans.
"Technology will never be a substitute for what people can do," she said. "We don't believe in that. Our culture [does not believe] in replacing humans with technology, but rather complementing them with AI tools so that it enhances the work we are already doing."
With this in mind, organizational leadership prioritizes operational efficiency improvements in the realms of length of stay, readmission rates, and clinical decision-making.
Viswanathan noted that, to date, BayCare has invested in pilots for voice-based AI assistants to help providers complete primary care visit summaries, generative AI chatbots for COVID-19 symptom triage, and sepsis identification technology, among other things.
The health system is also evaluating AI-driven tools that nudge patients to get necessary screenings or encourage them to ask questions before procedures.
BayCare's approach to assessing its efforts is patient-centered. Improvements in clinical outcomes are one of the organization's most important success metrics.
Viswanathan indicated that these improvements are a significant indicator that investing further into a solution or workflow might be beneficial.
However, she cautioned that stakeholders must weigh the pros and cons of how much a solution costs compared to its potential tangible outcomes. She explained that some tools -- like a sepsis indicator -- are "soft investments," because it can be difficult to determine just how much of an impact a tool might have on clinical outcomes prior to piloting it within the organization.
Stakeholders can also utilize improvements in operational efficiency and patient satisfaction to determine the success of a digital transformation initiative, Viswanathan noted.
"We want to make sure that we can get the patient experience and satisfaction scores," she said. "Is the customer satisfied with what we are doing? If they are not satisfied … then it will be a failure in our minds."
Finally, financial performance and cost savings are critical in helping health systems determine whether to pursue additional AI-driven digital transformation investments in a given area or wait until it is more cost-effective.
Alongside measuring these key performance indicators, taking advantage of strategic partnerships and having the right stakeholders in the room can make digital transformation less daunting for healthcare organizations.
"[Through partnerships,] we can monitor progress, gather feedback from people, and then adapt it to our strategies as needed," Viswanathan stated. "Engaging stakeholders and leveraging our partners to find out where the missteps have been [is key]."
"The other piece that I strongly advise is to have very clear, defined goals and outcomes that the organization wants to achieve," she continued, emphasizing that pilot projects for AI tools are time- and resource-intensive, so ensuring that a platform meets well-defined parameters can prevent health systems from pouring resources into pilots that are not likely to generate a return on investment.
Shania Kennedy has been covering news related to health IT and analytics since 2022.
