The majority of media attention concerning the use of artificial intelligence (AI) within healthcare has centered around exciting prospects such as robotic surgery, digital pathology, and next-generation radiology, to name a few. And yet delivering better care is not the only component of the Triple Aim in which AI can deliver significant value. New technologies incorporating natural language processing (NLP), machine learning, and other aspects of artificial intelligence are already making an impact to help improve population health and lower costs.
In this post, we’ll take a look at three important areas within healthcare that are likely to gain the most benefit from AI in the realistic future. Some of these gains may not be as exciting as those being used on the frontline of care, but it’s important to understand the ways in which AI (and its associated technologies like NLP and machine learning) can deliver a tremendous boost to health engagement, driving healthcare consumers to positive health action that results in longer, healthier lives.
Personalized Healthcare Communications
Although the purpose of CMS's Medicare Star Ratings is to improve overall health, initial attempts to improve quality ratings resulted in an overabundance of communication touchpoints. In some cases, members have reported receiving in excess of 200 messages from their health plan in a single year and some
studies report that experiences with payers and providers have actually
worsened in the past two years. Despite best intentions, this overabundance of mass communication has created a major problem known as “member abrasion,” damaging health engagement rates and even creating distrust of health plans in general.
AI can be of tremendous assistance in developing more effective health engagement that rebuilds trust and encourages positive health action. In a
2017 Deloitte survey, healthcare consumers reported overwhelmingly that they wanted to be “heard, understood and given clear direction tailored to their individual needs and preferences.” Industries like retail have proven that AI techniques can expand traditional data sets used to personalize communications, resulting in much higher rates of engagement. Similarly, “smart” health engagement technologies are emerging that capture critical variables around context, behavior, channel preferences and even values. This additional data helps drive increasingly personalized communications and provides better health outcomes in the long run. These systems use machine learning to measure responses to campaigns, replicate those that are successful in driving health action, and replace or remove those that are failing, often in real time.
When used successfully, AI actually “humanizes” engagement, resulting in empowering healthcare consumers to actively participate in their health based on their responses to personalized touchpoints sent with the appropriate frequency, and using the right channels.
Boosting Productivity and Reducing Costs
AI and machine learning solutions can often deliver a strong ROI when focused on administrative or marketing functions involving high volumes of data and a high number of transactions. For those reasons, it’s easy to understand why many companies with AI solutions target healthcare administration. The transmittal of data, scheduling, and many less critical tasks associated with prescriptions, insurance validation, and other processes are being streamlined by AI solutions. Not only can the use of AI streamline manual processes, it can also help reduce the burden of low value activities currently placed on many care providers, including physicians. One such use case employs NLP to help physician extenders and administrators synthesize rich unstructured clinical data from millions of pages of physician notes.
This past spring the Journal of the American Medical Association (JAMA)
reported that only slightly less than half of all expenditures in U.S. healthcare are spent on administrative functions including planning, regulatory, and management tasks. Perhaps a more mundane statistic (but one that still resonates) is that according to
Harvard Business Review, healthcare organizations still send over
120 million faxes per year. Rounding out that information is the fact that 14% of healthcare spending is wasted effort on poor administrative practices.
A new
study from Accenture recently reported that AI could help reduce costs within the healthcare industry by as $7 billion within the next 18 months alone. This could be accomplished by automating processes within areas like customer service, billing, enrollment, claims and quality and compliance. In the survey, Accenture consultants outlined six areas of operations where AI could drive significant cost savings for payers alone.
One organization that has embraced the potential of AI for administrative applications is KLAS, a healthcare IT and data insights research firm. KLAS has created the
ARCH Collaborative, an initiative involving over 150 health systems focused on improving the efficiency of EHR systems. KLAS continues to advocate for pursuing AI technology to help reduce unnecessary waste in administrative areas.
Gathering Insights from Wearables and Other Smart Devices
Improving overall population health requires the ability to gather significant amounts of data based on day to day behavior. Healthcare consumers are becoming increasingly adept at the use of wearables, virtual assistants and smartphone apps that track, monitor and report detailed health information. In fact, human activity is becoming one of the biggest sources of health information available to healthcare organizations.
AI can be used to expertly extract this data, creating smart segmentation and powering data analysis for pursuing community-based health strategies. In turn, this data analysis may be used for research to study the impacts of social determinants of health on overall population health, or to measure the success of ACO initiatives. Tools like big data and natural language processing are already in use to help not only screen for potentially fatal diseases like cancer, but also develop customized treatments.
For additional information on these topics and AI, visit our previous
blog.