Following other industries, there will come a day in the next decade when artificial intelligence (AI) is so nuanced and complex, that AI will be able to solve many of our health care system’s challenges. However, today is the day for AI assistance, as developments in AI technology for healthcare continue to mature. The health care community—and HIM in particular –needs solutions that can streamline and assist in routine tasks, allowing coding and auditing teams to focus on the more complex cases. AI technologies partnering with experts to increase productivity is empowering.
When technology solutions assist healthcare works, quality goes up, Today, despite our collective investments in health care software technology, it seems we are more focused on productivity versus quality. AI assistance combines productivity and quality.
While the current system is far from self-sufficient and complex Full AI might be just a bit beyond the horizon, there is a bridge between complex human intelligence and automated AI. And there is much to learn from other industries that are further along this innovation lifecycle.
Bridging the gap: Examples from the modern world
Economist Tyler Cowen wrote in Average is Over in 2013, “the productive worker and the smart machine are, in today’s labor markets, stronger complements than before.” It’s a trend we’re seeing across industries and there’s no reason to believe that health care will not also benefit from it. In the next few years, human workers will be made more efficient by technology solutions. We, in turn, will make machines smarter. And this virtuous cycle has the potential to elevate the health of our communities.
“AI can be more than an interpreter of human speech.”
In everyday life, AI solutions are becoming mainstream. Most modern smartphones come equipped with an AI assistant that easily translates spoken words into internet searches. The average smartphone user doesn’t even need to manually set an alarm to get up in the morning – all he or she needs to do is ask the AI assistant.
As we see in other industries, AI is more than an interpreter of human speech. In the automobile industry, cars on the market today have assisted steering to prevent collisions and help with difficult parking jobs. Tech giants such as Google and Tesla are currently developing fully automated self-driving cars – but even these advanced AI systems require the occasional input from a human driver.
A Google self-driving car may recognize that there’s a large truck in the next lane, and understand there’s a safe distance of two feet between the truck and the car. The human driver, however, feels that two feet isn’t enough space, so he nudges the car a little further away, until he feels more comfortable. Over time, the car learns that humans prefer a little extra room between themselves and large trucks and in the future makes driving adjustments accordingly. This is a clear example of how human intelligence is necessary to the improvement and advancement of AI.
Now consider a more personal example. Say a busy executive takes a sick day away from work and returns the next morning to an overflowing email inbox. Some of the messages are junk, others aren’t time sensitive, and a few need to be addressed right away. The executive can sort her inbox based on priority and quickly attend to the most important messages first. This prevents important emails from getting lost in the shuffle and makes the executive more productive.
Why not take it a step further? What if an AI assistant could automatically pick out which emails require an immediate response and queue them before the executive even logs onto her computer? She could then make her own decisions about which to tackle first based on her understanding of current trends in risk and company goals. A decade ago, this would have seemed farfetched, but here we are, in 2017, and Gmail does exactly that.
Streamline Health is applying these AI assistance principles to health care today. In coding, we are sifting through millions of code selections and rejections to learn how the coding process can be improved. This real data being used to assist coders and auditors better perform.
Building a bridge to better care
Hospitals should be able to focus on what they do best – managing the health of their communities. When the hospital’s revenue cycle runs smoothly, the revenue certainty allows for additional investments that benefit the community. When, the system is plagued with denials, longer DNFB, and allegations of overbilling, key resources are diverted to solve administrative problems that could have been invested in the community. Every minute a hospital executive isn’t playing defense on revenue cycle issues is a minute he or she could have invested in furthering the health of her community.
The fear-mongering of ICD-10 (those of us old enough can draw parallels to Y2K) fueled investments into computer-assisted coding, with the added hope that the process would become more efficient. In some respects it succeeded in doing so, but it’s far from certain that this will be the solution we use five years from now. Even in revenue cycle, we cannot be satisfied by plateaued productivity and quality. Our machines need to get smarter and we need to continue to build that bridge between the intelligence of a skilled HIM professional and the automated intelligence of a computer assistant. At Streamline Health, our eValuator™ does just that. eValuator empowers the smart working by adding smart assistance so that hospitals’ top auditing talent is only working on the most complex challenges and opportunities, improving overall accuracy and productivity.
The partnership between smart humans and smart machines will be powerful and transformative in health care as it has been elsewhere. The more routine tasks we can automate, the more quickly we can solve other non-routine problems. We’re excited, and inspired, by this vision every day.