Искусственный интеллект может снизить расходы американских клиник на программное обеспечение

$25 billion. That’s how much Forrester predicts U.S. healthcare providers will spend on software this year. That equates to approximately 36% of their overall IT budget. But the way providers are sizing up costs could be flawed. What if they are spending these substantial sums based on an enterprise software model that no longer makes sense? Providers absorb a lot of expenses associated with tasks like onboarding new employees, retraining staff, managing workflow fragmentation, billing overhead, and maintaining the administrative processes required to keep practices running. These costs far exceed what a typical solution’s price tag would lead you to believe. Practices buy into the narrative that enterprise software is expensive and will continue to grow more expensive. It’s just part of doing business, a necessary evil culture tells us. Yet, for the first time, we may see this narrative challenged. AI has the potential to address healthcare’s operational inefficiencies in a truly meaningful way. After spending years observing how software is bought, sold and deployed, I see that the math is about to change. Small Practices, Big Impact Take a medical office with five providers. These practices tend to maintain a lean staff, which means the prior authorization queue can back up, or it can take time to get an employee fully onboarded. Denied claims might take a billing specialist as much as two days to rework. All of the seemingly little things add up to become something big. Now think about what saving an hour a day per employee could mean. Across a team of 8-10 people, that quickly translates to recovering the equivalent of a full-time employee. This is not about letting staff go but rather saving time and overhead in a way that is fiscally responsible for a practice likely operating on thin margins; it frees existing resources to better serve patients and the practice. The same logic applies to denied claims. The average denial rate across independent practices runs between 5-10% of submitted claims, and many of those denials are preventable by closing documentation gaps and eliminating coding inconsistencies or problems with preauthorizations. If a practice has software that applies AI to catch and fix small details automatically, it can protect revenue that might take weeks to recover otherwise. There are examples across every stage of the workflow where a dose of AI would improve a practice’s overall health. Injecting AI into the processes that cause the most drag enables them to make real, structural changes in how their facilities operate. Reconsidering the Purchase Process It can be very exciting to imagine all of the things AI could accomplish within a practice, which in turn, could lead some practices to evaluate software purchasing decisions based on promises of capabilities. This would be a mistake, however. Organizations have built lengthy evaluation processes around functionality checklists and integration requirements. While those considerations remain relevant, they may not be sufficient. A more important framework for evaluation is one that focuses on reducing complexity. Platforms that eliminate workflow steps create more value than those that add a cool feature or two. This is because technology focused on simplifying operations impacts a practice at a deeper level, touching revenue cycle performance, staffing efficiency, patient and provider satisfaction, and more. Healthcare organizations are running in an environment where labor costs are rising amid a workforce shortage. Getting smarter with resources is imperative because most practices won’t be able to hire their way out. And part of getting smarter is changing what is expected of software vendors. Historically, healthtech pricing models expanded through implementation services, consulting, ongoing training programs, and user-based licensing. But as technology becomes easier to implement, learn and manage with AI, these add-ons are no longer necessary. Additionally, AI is compressing the cost of software production itself. Solutions are being developed faster and cheaper than ever before. As such, practices should no longer be forced to pay an ever-increasing premium. And with AI’s ability to successfully onboard new users quickly, practices won’t be afraid to switch to another software provider if they aren’t satisfied. The barrier is far lower. Questions We Should Be Asking Instead of buying software the way we always have, evaluation and purchase criteria need to evolve. We should be asking questions like: Does the solution reduce complexity? Does it decrease operational inefficiencies and simplify onboarding? If so, how? Does AI materially reduce engineering overhead? If it does, why should my costs go up? (They shouldn’t.) Ultimately, the conversation moves from “What does this software do?” to “What does this software eliminate?” This creates a buying framework built around different types of vendor conversations. Enterprise software spent two decades getting more expensive and complex. Healthcare software followed suit but then layered a sophisticated regulatory and billing environment on top. Providers absorbed it for lack of a better alternative. The better alternative is here, thanks to AI. Both practices and vendors should take note. About Venky Chellappa, Ph.D. Venky Chellappa is vice president of strategy at CharmHealth, a leader in healthcare technology solutions for providers. In this capacity, he has grown CharmHealth from a startup venture to one of the leading EHR vendors, offering fully integrated solutions to the ambulatory and healthcare industries. In addition, Venky serves as one of the managing partners of the CharmHealth+Bioverge Digital Health Transformation Fund. His goal is to bring innovative ideas to the point of care along with CharmHealth.