Also known as machine learning or machine intelligence, artificial intelligence (AI) is a category of computer science that mimics the thought processes of the human brain. While difficult to concisely define and explain, it is multidisciplinary in both theory and practice, and has the potential to considerably improve quality of lives through its many applications. In medicine, AI is most likely to be utilized in diagnostics, risk prediction, treatments, drug research and clinical care. Not surprisingly, the rate of AI’s acceleration in application in medicine has been impressive, particularly among cognitive computing and big data analytics.
The Emerging Roles of AI and Precision Medicine in Heart Disease
One of the biggest challenges in managing heart disease is predicting which patients may be likely to experience heart failure as well as respond to available treatments, including medications. Artificial intelligence (AI) has become increasingly useful in detecting and predicting cardiac disease and heart failure, when based on accurate medical records and other data in predicting potential outcomes. Through clinical trials utilizing AI and precision medicine, researchers and providers are more able to pinpoint at-risk patients, with the goal of predicting and possibly preventing a cardiac event.
For patients, AI applications include medication reminders, remote follow-up, real-time medical counseling and detection of early warning symptoms. From a clinician aspect, AI can help with data collection, such as using voice information for taking medical information and connect providers with multiple health records systems while reducing physician workload. With AI, clinicians will be better able to implement a more accurate, personalized patient medical plan in the not-so-distant future, as cognitive computers can provide data to help them make better decisions and predict outcomes.
Research from Dawes TJW suggested that AI could predict mortality rates for heart disease patients. AI-based software recorded the results of MRI scans as well as blood tests for 256 heart disease patients, mapping 30,000 points marking heart structures in each heartbeat. Combined with patient eight-year medical records, data was able to predict the conditions leading to patient death, as well as predict five-year survival rates, with prediction accuracy reaching 80 percent, compared to the clinicians’ 60 percent accuracy rates.
Structured and Unstructured AI Data
Structured data, the basis of most current research, is labeled organized into formatted fields. A study was undertaken using structured data concerning the ability of machine learning algorithms to accurately predict five-year mortality rates among 10,030 patients with coronary artery disease, analyzing 69 parameters in the set.
The researchers’ findings showed that the five-year mortality prediction with the machine learning algorithm was measurably better than other traditional risk scores.
Unstructured data includes imaging analysis, textual EHR information and other data stored without a well-organized structure or formatting. AI is anticipated to open up avenues to utilize the wealth of patient data¸ estimated to be 80 to 90 percent, to patient research. Cardiac imaging analysis includes optical coherence tomography, cardiac single-photon emission computed tomography, MRI and intravascular ultrasound are additional AI applications, with machine learning providing more accurate and faster results in the future.
Diversity, AI and Precision Medicine
Precision medicine’s ultimate goal is to move away from a “one-size-fits-all” approach to that of more personalized patient care, tailored to the individual’s particular needs¸ including metabolism, genetic influences, and environmental and sociological factors. According to information from the journal Circulation: Genomic and Precision Medicine published by the American Heart Association (AHA), precision medicine in tandem with AI could some day be used in personalizing heart failure diagnostics and therapies by identifying high-risk patients and contributing to significantly improving heart failure care.
Research is showing that certain drugs are effective for treating heart failure, with resulting improved prognosis. However, there are patients who either experience adverse side effects or who don’t respond at all to treatments. Among the most important outcomes of AI has been the realization that gender, age and racial as well as socio-economic differences can account for why some treatments and medications work for some groups, such as white males of European ancestry, and not for females or people of other races and genetic makeup.
In the past, most participants have been of European ancestry, as well as predominantly male. Having finally become aware of the roles played by diversity, researchers are addressing these issues by increasing the diversity of clinical trial participants to find the optimum treatment protocol for a given population group. Gender, socio-economic factors including access to health care, nutrition and individual genetics now factor into AI medical research.
Researchers are backed up by the FDA, which recently released guidelines to help improve clinical enrollment diversity for trial sponsors. In 2017, they issued a report confirming that the majority of clinical trials fell short of real-world demographics, with African Americans making up only seven percent of participants. Women overall fared somewhat better, with 48 percent, but there is still much room for improvement.
With more realistic trial populations, the potential greatly has increased for developing more advanced forms of treatment, including medications reflecting biomarkers. However, there is still much research to be done for progress with other forms of precision medicine to enable them to transform treatments for heart failure. As with any technology, AI has its limitations of AI including potentially misleading results due to inaccurate or poor quality data selection.
Partner with an experienced medical billing company for cardiac claims
Billing for cardiac and related health issues presents a challenge for physicians treating heart disease. New technologies and treatments traditionally meet with payer obstacles to reimbursement. An experienced medical billing partner such as M-Scribe can guide your practice through the complexities of the billing process, ensuring that your claims are complete and meet payer as well as state and federal guidelines and requirements. Since 2002, we have helped practices of all sizes and specialties with their billing and practice management needs. Contact us at 770-666-0470 or email to learn how M-Scribe can help your practice increase and manage your revenue cycle.