The coronavirus pandemic has not just put the world's healthcare in intense pressure but it has also exposed the shortfall. There is a gigantic shortage of clinicians, nurses, medical supplies, hospital beds, and other medical infrastructure. All elective surgeries and usual care are considered as non-essential and are not getting the required attention. This has brought double whammy for patients with chronic diseases. First they are not getting any care and second they are also most vulnerable of getting infected with Corona. As COVID-19 has forced billions of people to stay lockdown in their home, the only plausible solution for chronic patients is to get remote and digital care.
Chronic diseases (such as heart failure, hypertension, diabetes) take up the majority of healthcare spending, accounting for 17% of the United States GDP. The challenge is too many patients and too few physicians. Other industries are solving this problem by using structured algorithms to make product development more scalable, efficient, and with minimized variation in quality. Can healthcare do the same to make patient care more algorithmic, scalable, cost-effective, while optimizing patient outcomes and reducing its variability?
Chronic Disease Management (CDM) is an integrated care approach to manage chronic diseases or conditions that are persistent or otherwise long-lasting in its effects. According to the US Centers for Disease Control and Prevention, chronic conditions such as heart disease, asthma, arthritis, diabetes, and chronic obstructive pulmonary disease (COPD) affect about half of all adults. Diabetes and heart disease are two of the costliest and most prevalent chronic conditions impacting patients in the US, leading the healthcare industry to spend billions every year to treat and manage these disorders. Some of the top challenges in managing chronic disease, apart from high healthcare cost, are screenings, check-ups, monitoring and coordinating treatment, and patient education. Nothing can replace clinician judgment in patient care, but new data-driven technologies such as AI are helping advance delivery of the right care, to the right patient, at the right time. As care delivery continues to evolve from reactive disease treatment to proactive, preventive care, more organizations are looking toward digitization in healthcare with advanced technologies such as artificial intelligence and machine learning to assist with drawing actionable conclusions from their big data resources.
AI has already institutionalized in other sectors
AI has proved an increase in productivity and growth in various sectors, including manufacturing, auto, and retail and e-commerce. It can increase the efficiency with which things are done, vastly improving the decision-making process by analyzing large amounts of data. AI can be very helpful in providing an integrated care approach required to manage chronic conditions. Recent advancement in image processing, deep learning, neural network, and NLP has opened the way for new possibilities that were unimaginable till now.
Recently, scientists at Google developed a new AI algorithm to predict heart disease by analyzing retina scans of a patient’s eye. The company’s software is able to accurately deduce data, including an individual’s age, blood pressure, and whether or not they smoke. This can then be used to predict their risk of suffering a major cardiac event—such as a heart attack—with roughly the same accuracy as current leading methods. This shows that possibilities are unlimited, and we are still scratching the surface of what can be achieved by AI.
AI can help in streamlining care for chronic diseases
A study published in Harvard Business review (https://hbr.org/2017/05/how-machine-learning-is-helping-us-predict-heart-disease-and-diabetes) explained how Paschalidis and his team in Australia worked on a project in 2017, in which they used patients’ electronic health records (EHRs) and machine learning to predict hospitalizations due to diabetes and heart disease. Using this method, the group found that they could predict hospitalizations about a year in advance, with an accuracy rate of up to 82%. Now, Paschalidis and his team will develop even more comprehensive predictive capabilities using EHRs and real-time health data, including information from wearables, implantable devices, and home-based networked diagnostic devices.
This shows that it is possible to develop machine learning algorithms that can identify patients at higher risks of heart disease or diabetes. Care provider can use these algorithms to enable early interventions and personalized treatments for high-risk patients.
Figure-1 CDM Journey
CDM journey starts with diagnosis of disease that can be triggered by patient reported symptoms or routine check-ups. A treatment plan is then created for the patient by the clinician. Patients adhere to the plan and keep monitoring for effectiveness. The plan can be updated by the clinician if, required on assessment.
Recent research shows that AI technology is just as good at diagnosing illness as humans. Specifically detecting disease, based on medical imagery and ECG data, when coupled with patient demography and health history. Image analysis and the ML model can study historical medical data (from other patients) to recognize a pattern similar to other patients to detect a disease. These ML models need to be ‘trained’ through data that are generated from clinical activities, such as screening, diagnosis, treatment assignment, and so on, so that they can learn similar groups of subjects, associations between subject features and outcomes of interest.
According to an article by Jiang on National Institute of Health (NIH), there are several types of AI techniques that have been used for a variety of different diseases. Some of these techniques discussed by Jiang include: support vector machines, neural networks, decision trees, and many more.
Figure 2: Road map from clinical data generation to clinical decision making (Source: National Institute of Health (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5829945/)
The flow chart in figure-2 describes the road map from clinical data generation, through NLP data enrichment and ML data analysis, to clinical decision making.
Though more qualified studies are required to substantiate the research, this is a significant progress in providing scalable, fast, and cost-effective diagnosis of a chronic disease.
In a recent experiment, IBM Watson has made a promising progress in oncology. According to NIH, in a cancer research, 99% of the treatment recommendations from Watson are coherent with the physician decisions. Furthermore, Watson collaborated with Quest Diagnostics to offer the AI Genetic Diagnostic Analysis. In addition, the system started to make impact on actual clinical practices. For example, through analyzing genetic data, Watson successfully identified the rare secondary leukemia caused by myelodysplastic syndromes in Japan.
The sophisticated algorithms and ML models can be trained through healthcare data and the system can assist physicians with medication modelling and treatment suggestions. These models suggest appropriate dosage and treatment plans by analyzing practice guidelines, meta-analyses, patient characteristics, and clinical trials data.
AI can also help in creating personalized plan for individual patients instead of a one-size-fits-all treatment, assessing the stage of disease rather than the individual patient. The personalized treatment helps the clinician to intervene before a patient’s condition becomes critical, resulting in lower cost and better care.
Care management is usually a lifelong process in case of chronic diseases, mostly self-managed by the patient. A virtual nurse assistant based on AI can help patients in ensuring drug adherence and monitoring by recording, reminding, and connecting with biometric devices to collect vital data.
Recently, Sensely, a meditech start-up, introduced AI-powered nurse avatars. These avatars cover a wide range of content modules, including system assessment, health information, and wellness and chronic care, and are available in over 30 languages. Each morning, patients receive a notification to complete a check-in routine, with Sensely avatar ‘Molly’ guiding patients to record weight and blood pressure data. Sensely then calculates a risk assessment for each patient and provides clinicians with timely information to trigger appropriate interventions.
These types of virtual assistants are very convenient for patients to self-manage chronic conditions.
Another critical area in care management is measuring and managing chronic pain. Sometimes, patients live with chronic pain that is often debilitating and hard to diagnose clinically. AI and facial recognition algorithm can assist in detecting and monitoring chronic pain in patients by monitoring facial muscle movements in individuals who can’t self-report pain, providing a quantifiable score of the pain they’re experiencing. This information is very critical for clinicians to adjust the treatment plan.
Apart from helping patients once they are diagnosed with a chronic disease, AI technology can even do early detection of disease signs in patients before they become chronic. AI-based ML models can help providers identify patients who are at risk for heart disease, hypertension, and pre-diabetes, allowing for early intervention and preventive care strategies. For example, a retinal examination by artificial intelligence may reveal a person's risk for heart attack or stroke in the next five years, according to a new study by Google.
AI can also be crucial in preventing hospitalization of existing chronic patients who are going through the treatment. Continuous monitoring of patient vitals along with drug adherence by an AI-based system can detect the possibility of deteriorating conditions, thus needing hospitalization.
With the onset of digital health management where data is being collected from various touchpoints throughout the patient lifecycle—including wearables, mobile apps, and hospital’s EHR system—we have tremendous amounts of data on patient activity, patient-reported outcomes, patient history and treatment plan. This data can generate new insights and possibilities for affordable and scalable care with the help of AI applications. Through machine learning, this data can be harnessed to prioritize patients based on real-time needs, generate intervention alerts, and recommend follow-up actions.
AI, along with sensor technology, can unlock the potential of data, providing actionable insights to guide clinical decisions to diagnose, treat and manage chronic conditions remotely, thus, providing the right clinical intervention to the right patient at the right time.