In this FiveMinuteFriday episode of the SuperDataScience Podcast, we focus on how data science can help advance the healthcare industry. Billions of people can benefit when a more focused integration happens. Start tuning in to know more!
Data Science, Artificial Intelligence, and Machine Learning can surely provide the needed solutions for the healthcare industry. These emerging technologies have already proven their value and have defined industries like communications, banking, retail, etc.
So, today, we discuss how data science can change the healthcare industry for the better. It doesn’t matter if you’re in a different sector since there will be a thing or two that you can take from this episode and apply to your career. This also broadens your horizons and serves as a good starter if you want to jump into the data science and healthcare industries.
I’ve consolidated all the benefits that can happen if data science is integrated with all healthcare systems and operations. I’ve managed to narrow it down to these 5 distinct applications on healthcare:
- Improving diagnostic accuracy and efficiency.
- 5% of adult patients are misdiagnosed every year in the US alone. That’s almost 12 million people that we’re talking about. And, data science, Ai, and machine learning can significantly lessen this number.
- Using wearables data to monitor and prevent health problems.
- There are already wearable and implanted devices that are used to monitor our bodily functions. The collected data can then be used to create a medication, diet, or lifestyle to maintain a healthy body and prevent chronic illnesses in the future.
- Genetics and genomics.
- It’s getting cheaper and cheaper to encode a human genome. From the data collected from you, your family, or ancestors, you can easily see if you’re susceptible to hereditary defects or illnesses.
- Creation of Drugs.
- It takes too long to get a drug out in the market and be availed by those who need it. It delays treatment and recovery. Data science can help hasten the process.
- Virtual assistance for patients and customer support.
- We’re not new to the fact that the population is still growing. There may come a time that there will be no enough medical staff to respond when you need it. AI can help give assistance if it’s only a minor issue.
So, there you go. I’m pretty sure there’s more that emerging technologies can put on the table so let me know if you have something to add!
- IBM Watson
- SDS 013: 95% Accuracy Models, Winning People Over, and Saving Lives with Damian Mingle
DID YOU ENJOY THE PODCAST?
- What do you think is the most significant application of data science in the healthcare industry so far?
- Download The Transcript
- Music Credit: Candyland by Tobu [NCS Release]
This is FiveMinuteFriday, episode number 216, Data Science in Healthcare.
Welcome back to the show, ladies and gentlemen, super excited to have you here today, and this time we're looking at an industry application of data science, specifically we're talking about healthcare. Now the way we're going to go about this is we're going to look at five distinct applications of data science in the industry of healthcare. The purpose of today's podcast is to broaden our horizons. Whether you're already working in the space of healthcare or you're working in a completely unrelated industry, or maybe you haven't yet decided which industry you want to work in, this information will help you broaden your horizon and look at different problems and challenges which data science is able to solve.
The beauty about data science is whichever industry applies in, data science is data science. These skills are highly transferable, so by learning something from a completely unrelated industry, you can actually come up with better ideas, be inspired to look into different methodologies in your own work, in your own industry. Let alone, of course, if you're already in the space of healthcare then some of these applications might be very beneficial for your career as they are. On that note, let's not put it off and without further ado, we'll dive straight into it.
Industry application, number one. Improving diagnostic accuracy and efficiency. This is quite a straightforward and obvious application of data science where we can help by improving how accurate our diagnoses are, but before we talk about specific use cases or specific examples, let's talk about the facts. About five percent of adult patients every year are misdiagnosed in the US and that is a total of over 12 million people. 12 million people receive the wrong diagnosis every single year, and that's just in the US alone.
Moreover, what is an even more important probably sad fact, very sad fact is that postmortem analysis or examination shows that about 10% of patient deaths happen because of misdiagnosis. That's 10% of patient death happen because of the wrong diagnosis being issued. As we can see, that's a very large problem, it should be addressed and if data science can help in any way, it should.
There is plenty of ways that data science is already helping with the situation. We've probably all heard the example is how IMB's Watson, the same machine that beat humans in Jeopardy, how it was able to diagnose a woman with a rare form of Leukemia, so that woman came to a hospital in Japan and they diagnosed her with a certain type of cancer. They treated her with chemotherapy, and then she wasn't recovering as fast as they expected, so they went to IBM Watson, and the machine diagnosed her with a different type of cancer, a rare form of Leukemia and that was the correct diagnosis. The initial diagnosis was incorrect and the treatment was actually not helping.
In that case, IBM Watson was able to help, as you can imagine, that happened back in 2015, in January 2015. You can only imagine how far we've progressed from there and how much more powerful machines are and how much better they can help with situations like this. There's plenty, plenty of other examples, on the podcast we had a guest at the very beginning of the show, on episode number 13, Damian Mingle, and he was talking about a project that he was working on where patients would come into emergency rooms and just through computer vision and analysis of their heat patterns and how they looked and some basic tests that they did upon coming in, the machines could detect whether they had a fatal, a very lethal disease or not, that they needed to be treated as a priority or they could wait in the emergency room for a little while. There's plenty obligations and that's just one of the use case of data science in healthcare improving diagnostic accuracy and efficiency.
Let's move onto to number two. Using wearables for monitor and prevent health problems. Once again, let's talk about the fact the every single year about 600,000 people suffer sudden heart stoppages in the US. Astronomical number and that's just one of the examples of things that might happen to a person on the street. What if we could prevent that? That's where wearable devices that we can carry on us all the time, for instance, like a watch or something that monitors your heart rate and heartbeat and things like that. That is where we can use their data to help prevent these situations and help alert the person or emergency services when something like that is about to happen or might happen.
That also goes for habits, for sleeping habits, for eating habits, for exercise habits, all of those things, how we live our lives and how we treat our bodies, those things can lead to chronic illnesses and lethal outcomes. Wearables can help prevent those situations at the very start and that is exactly how data science can help not just diagnose the problem but prevent the problem from happening in the first place.
Use case number three: genetics and genomics. As we progress in the 21st century, it is becoming cheaper and cheaper to get to encode the whole human genome for a person. Alright, now it costs about $600 and takes a couple of weeks to get it fully encoded and from that information you can extract so much data. You can extract data on diseases that you might have, on things that might effect you in the future, about your ... you can even extract data about your ancestry, where potentially which parts of the world your ancestors came from. You can extract any recessive gene mutations that you don't ... that are not prevalent in you, that are not being ... that aren't demonstrated or not taking effect in you but might have affect in your children. There's lots of data that you can extract from there, and that is a huge green field for data science.
There are some applications that are coming up, for instance using data science to analyze what kind of genes or gene mutations don't go well with certain types of medicine or medications and other things like that, but those are just starting points. This is a whole, huger field where data science will have more and more power, especially as geo-coding, or especially as coding the genome is going to cheaper and faster and more people are going to jump onboard.
There will be services starting to pop up on how data science is helping people get valuable insights from the information, which is contained within their gene.
Use case number four: creation of drugs. The fact is that on average is takes about 12 years to get a drug officially submitted. It's a very long time and a lot of people can't wait that long or we improve the lives and health of many people if we get these drugs submitted faster, but still, in an ethical and at the same time safe and fully tested way. That's where data science and machine learning come in, they help simplify this process because we can actually do a lot of the assessments that take time with machine learning and data science we can do tests and we can do simulations, we can find out what kind of success rates to expect based on certain biological factors and basically speed up the process incredibly to cut down those 12 years to a couple of years. So that drugs can get out into the world and start helping people faster than before.
Data science, once again, can help there because we have plenty of data on how other drugs work and we could compare the new drug to previous cases of a drug or run simulations and tests and perform statistical significance tests and things like that to make sure that the results that we come up with are indeed valid.
Finally, use case number five: virtual assistance for patients and customer support. More and more we are seeing chat bots pop up and artificial intelligence powered support systems pop up in order to help people because ultimately, population is growing, and in many cases there's just not enough staff to help everybody. How many times have you been to emergency room and you had to wait for several hours, like three hours or six hours. What if you issues could be solved by artificial intelligence faster?
What if you could have an app with you that could help you resolve some of your questions or help you assess certain things. For instance, you might have an injury or maybe you want to look at a certain spot on your skin, you want to assess that. Instead of going all the way to a doctor, getting an appointment, waiting, you could just download an app, take a photo and within seconds have an answer. Like whether it's dangerous or it's not. Dangerous, potentially, benign or it's a concern that you have to go and investigate further. That could cut down a lot of time, cut down a lot of cost and really help prioritize or help get those people through that really do need help, that have severe cases that need urgent addressing, get them through faster. In essence, improving everybody's experience. Not putting anybody on hold, not putting anybody at a disadvantage.
That's just another example of how data science can help and that's where we would apply things like natural language processing, artificial intelligence, deep learning and other technologies like that.
There we go. Those were five very powerful applications of data science in healthcare. They're plenty, plenty more. We're not going to go through all of them on this podcast, but I hope that even those five were valuable to you and you got some interesting ideas or interesting insights from them.
We'll continue looking at other industry applications in some of our other FiveMinuteFriday episodes, so stay tuned. If you'd like to get some links for the resources that we used in this episode, head on over to SuperDataScience.com/216 and there you'll be able to find out more information on data science in healthcare.
On that note, thank you so much for being here today, I look forward to seeing you back here next time, until then, happy analyzing.