For the last century, advancements in medical science and technology has been extending life expectancy's significantly. As of 2016, figures released by the World Health Organization (WHO), estimated the average global life expectancy was 70.5 years, with the life expectancy in US slightly below average at 69.1 years.
This means that a larger percentage than ever before of the US population is aging and this trend is only going to continue. The Population Reference Bureau predicted that the number of Americans over the age of 65 will double from 46 million people today to 98 million by 2060.
This has pushed a lot of the conversations away from just increasing the length of lives to how we can utilize all this knowledge and technology to improve ensure our quality of life once we get there. We spoke to David Kho MD, Chief Medical Informatics Officer for Chenmed LLC to ask him how tech is going to care for us as more of us enter our golden years for longer.
Innovation Enterprise: With an aging population, how important will wearables become to healthcare?
David Kho: There are 3 reasons for why wearables will become increasingly vital to healthcare:
1. Early warning - By detecting early warning signs of an impending, serious medical event, timely steps can be taken to intervene and help the patient before things escalate. The classic example is of a wearable or implantable cardiac monitoring device detecting dangerous heart rhythms and alerting both the patient and provider of an impending heart attack. However, there are other less obvious examples, a wearable or ingestible device with an accelerometer measuring changes in physical activity can indicate changes in sleep patterns, mobility ability, energy levels, etc.
2. Patient empowerment - Healthy behaviors are ultimately what will achieve the best outcomes. By providing data about their health directly to the patient, he/she is empowered to make the health-promoting decisions in real-time. Numerous studies have also shown the effectiveness of biofeedback control of vital signs when a patient is able to view his/her vital signs.
3. Improved Medical research - With the appropriate protections and controls for patient privacy and security, aggregated data collected from wearables will transform the way we conduct clinical trials. Currently, many clinical trials suffer from a lack of diversification in their study subjects, this affects how the FDA and ultimately a doctor decides on the dosing and effectiveness for a specific patient who may not have been represented in the original drug trial. Wearables allow for the collection of real-world data from many more to augment the formal clinical trials.
IE: Do you think digital literacy will become integral to geriatric healthcare?
DK: Surprisingly, no. Human factors design in technology has removed much of this barrier, making digital literacy less relevant as the machine interactions become more human. Consider this example: a research think tank in California placed virtual voice assistant devices (Alexa) in nursing home residents. Over 71% of the seniors subsequently report using the devices regularly, and part of the success was reported as due to the fact that interactions with the virtual voice assistants are so natural and intuitive. Effective technology will learn about us and our way of communication, instead of the other way around.
IE: Do you envision a future where some kind of data-based degree becomes essential to becoming a medical doctor?
DK: The standard medical school curriculum already includes extensive required material on advanced statistics. Traditionally, this was undertaken from the perspective of being able to understand clinical trials or public health studies. Statistics form the basis of Data Science and Data Analytics. I anticipate that the depth and complexity required (as well as the reliance on data science and analytics technology and software) will only continue to grow, you see this in other fields as myriad as economics, astronomy, physics. Essentially, what has happened is that the data science and analytics have become a core curriculum (like Algebra or English) and the problem-solving focus becomes the specialization (healthcare, economics, physics, etc).
IE: …So going forward, how important will patient analytics, machine learning and artificial intelligence become to the future of healthcare?
DK: Patient analytics occupies a very special place in data science for me. Basically, It can be broken down into 2 conversations. One is about patient predictive analytics. The other is about patient prescriptive analytics.
Let me take you through patient predictive analytics first. By predictive, we mean of course forecasting, and in this case, we want to forecast what the likely outcomes our patients will have given what we know about them. This is tremendously useful for risk segmentation. Patients can be sorted and ranked on a list based on the predictive risk of an adverse event occurring, for example, the likelihood of ending up in the hospital. For the patients at the highest risk, we can assign expensive and limited resources such as care coordinators, home health nurses, house call physicians to them, since we know that they have the highest needs. It would not be feasible to assign these resources to everyone due to cost and limited availability. So patient predictive analytics allows us to focus our efforts on the patients who need it the most.
Patient prescriptive analytics is entirely different. By prescriptive, we mean setting up data-driven rules to decide on what is the right course of action or intervention to avoid a bad outcome. In the context of health-care, this has very strong implications for preventative medicine. A simple example of a rule-based decision tree is establishing and then avoiding adverse drug effects. For example, it has been shown that ace-inhibitors will induce severe coughs in individuals of Asian descent, therefore many physicians will seek to prescribe something else for these patients to increase medication adherence and decrease the risk of aspiration pneumonia. This works well for previously established/known adverse events. But what about ones that have not been reported in the literature? This is where AI and Deep Learning comes into play. Machine learning excels at pattern recognition. In the same way that a well-trained neural net can recognize images of a cat vs. dog in ways that we cannot articulate, the same neural net can also be used to uncover hidden features of a patient's disease pathway that we don't yet understand. Those hidden features can help us gain additional knowledge and insight. This can then become useful for us to better understand the pathophysiology (ie what is causing the disease) and also helps us better tailor the treatment specific to each patient.
To find out more on how you can leverage data analytics to deliver patient-centric outcomes, visit this year AI & Data for Pharma Summit, part of DATAx New York on December 12–13, 2018.
Book your place HERE.