Artificial Intelligence is transforming the biomedical field. In particular, new data analytic methods can augment clinicians’ ability to interpret and act on large and complex data sets. While healthcare practitioners cannot yet be completely autonomous, technology plays a major role today in improving outcomes, increasing safety and reducing healthcare costs.
Globally, the healthcare industry is expected to increase AI investments by 44 percent between 2017 and 2020 (Tata Consultancy Services survey). Healthcare-focused artificial intelligence tools are expected to bring in more than $20 billion by the middle of the 2020s, exhibiting a compound annual growth rate (CAGR) pegged at anywhere from 40 to nearly 50 percent over the next five to seven years.
And, in the first nine months of 2017 alone, global healthcare big data analytics companies received more than $1 billion in venture capital investment, representing a 31 percent increase over funding levels at the same time in 2016 (Mercom Capital).
In this frenetic environment, life science leaders must balance their need to apply these technologies to improving patient care with bottom lines – without getting caught up in the healthcare AI hype machine. Healthcare leaders must ask themselves: is it worth investing in tools and infrastructure before seeing proof of their potential? And if so, how can prospective enterprise customers balance caution with the reality that machine learning will be the next transformative force in healthcare?
At the basic level, we see a trend in clinical decision support (CDS) which offers clinicians, staff, patients and other allied healthcare providers with personalized information, at specified times, which is designed to enhance health outcomes. Support devices such as reminders, alerts and evidence links to electronic medical records will guide the clinician’s decision-making. In this clinical space, physicians and practice managers should apply validated technology from companies which offer AI software to assist physicians with disease diagnosis and treatment algorithms. The development and use of clinical decision support systems should be driven by a core framework that articulates appropriate conditions for their use. These include processes for monitoring data quality and developing and validating algorithms, and sufficient protection of patients’ data.
In the biopharma industry researchers will be utilizing AI and machine learning technology to drive decision-making processes for existing drugs and expanded treatments for other conditions. Pharma leaders will also employ AI tools to expedite the clinical trials process by identifying patients from various data sources. Pharma is also using AI to predict the time, location and likelihood of epidemic outbreaks. Services such as Deep 6 Trial AI and IBM Watson for Clinical Trial Matching are providing tools to find eligible patients for complex clinical trials within minutes, depending upon the criteria. These tools, however, must allow for human validation and training of the AI software in order to resolve any data inconsistencies or conflicts.
Large primary care institutions are utilizing AI technology to reduce hospital readmission and prevent medical errors. Ultimately, hospitals hope that the application of AI will help identify and prevent high-risk patients from developing complications and guide patient care decisions. Hospitals will also utilize AI to streamline workflow and reduce or eliminate procedure redundancies, saving precious time and limited resources. Hospital customers should proceed with caution when considering the capital expenditure for AI technology. An algorithm may work in an academic or limited clinical setting, but it may not translate into the real-world community hospital. If an AI tool is trained by using data from a research hospital, it may not function well in a regular hospital where many patients have incomplete medical records. Healthcare organizations must ensure that their AI tools have been trained and validated with representative populations.
Currently a third of all healthcare AI Software as a Service (SaaS) companies are focusing partly or exclusively on diagnostics – making it one of the largest focus areas for burgeoning startups in the field. The Institute of Medicine at the National Academies of Science, Engineering and Medicine reports that “diagnostic errors contribute to approximately 10 percent of patient deaths,” and also account for 6 to 17 percent of hospital complications.
At Guerbet we are utilizing AI applications to ultimately improve the diagnosis and treatment of patients. As a global leader in medical imaging, we have identified AI partners with validated software technology and solutions. For example, we have partnered with IBM Watson Health to develop artificial intelligence (AI) software to support liver cancer diagnostics, utilizing CT and MRI imaging, and care.
Currently, AI in healthcare finds itself in the second stage of the Gartner Hype Cycle – “the peak of inflated expectation.” However, if we do not allow this vital complement to healthcare delivery to catch up to the hype, it may fall back into what Gartner calls the “trough of disillusionment.” To avoid this, it is important to properly understand AI and its limitations so that we may use it more effectively in areas where AI particularly excels.
And, arguably, ensuring the improvement, extension and quality of human health is one of the most valuable applications of AI. Life science professionals must continually balance the cost of AI technology with its value.