Welcome to episode nine of the new Ausmed Handover podcast: Doctor Google has Arrived: What does this Mean for the Nursing Profession?
Welcome to episode nine of the Ausmed Handover podcast. The internet- as we know it – has been around for over three decades now, and if you are a ‘user’, then companies such as Microsoft, Facebook, Amazon and Google have accumulated an extraordinary amount of very personal information about you. So much so that in early 2016 Microsoft, Google and an independent medical team all claimed that by analysing very large amounts of global browsing data they were able to accurately diagnose serious diseases in specific individuals. What are the ethical issues arising from this, and what are the implications for nursing practice?
Hello, this is the Ausmed handover podcast, my name is Darren Wake and in this episode, I want to examine how Doctor Google has now changed healthcare forever, and what it means for nurses.
DIGITAL SOUNDS (0.05)
The internet has been part of our lives now since about 1990, and these days, several hundred million computers are interconnected via the world wide web, plus around six and a half billion mobile phones, over 10 billion peripheral devices and probably twice this number of lesser machines that either relay or receive operational data via the internet.
Continue Reading Transcript
We now use the internet every day to communicate with one another, manage our finances, conduct business, work, purchase goods, educate and entertain ourselves, navigate our cities and for countless other reasons.
It is undoubtedly a permanent part of our lives, and in many, many ways, it’s a one-way trip: so many aspects of our day to day existence are now so entirely dependent on the integrity and continued function of the internet that if it were to cease functioning, society as we know it would collapse into chaos.
The core purpose of the internet is to transmit data, and transmit it instantaneously. If we just restrict our concerns to your average everyday schmuck like me, then this has countless practical applications: shopping, banking and personal communications with friends are some of the things that I use the internet for, along with the even more common use of adding to my personal database of knowledge. If I need to know something, then I simply type a question or a keyword into the search bar of Google or Bing, and in less than a second, I literally have several thousand links returned, all ranked by the search engine according to what it perceives is their relevance to me.
So, it seems that the internet enhances our lives in a lot of incredible ways, but how we use the internet, those seemingly trivial every day exchanges and requests for information, has been of profound interest to governments, spy agencies and especially big business pretty much since the day the internet went public.
Although there were many search engines around in the early days of the internet, in the late 1990’s Google rose to dominate the industry and pretty much wiped the floor of its competitors simply because the search algorithm it used ranked websites according to their relevance to the search term entered by the end user on their browser.
End users are people like me, and you.
Now this seems all very nice and beneficial, but the two chaps that developed Google, Larry Page and Sergey Brin, weren’t so much interested in helping you find the information you wanted to know as understanding your behaviour as a consumer and collating a very comprehensive demographic profile about you and your online behaviour.
And how they do this is very, very interesting.
DIGITAL SOUNDS (0.05)
Every day, every second really, Google’s servers index millions of web pages all over the world. How it does this is fairly techy and complex, and involves millions of tiny internet programs that crawl the internet on a continuous basis, but what they are seeking to do is build up a list of keywords and search terms related to individual sites so they can be found by end users who enter these search terms and key words into the google query bar on their internet browser.
However, that’s not much of a business model: helping people find information they want is a service Google offers, that’s true, but Google isn’t really a librarian, it’s a multi-billion-dollar business that understands with absolute clarity that the real and substantial profits lay in collecting a wealth of information about you and especially your online behaviour, because that can be converted into profit.
Google (and Microsoft’s Bing, social media sites such as Facebook and some of the other remaining search engines) are able to collect very detailed information about you every time you log on, go online, whenever you navigate your way through any website and what you do on that website when you are there.
This information is incredibly unique and extremely valuable and over times it’s possible to build up a very accurate picture of your online behaviour.
When you are online, subtle Google programs are literally walking through the contents of your computer, collecting information about your contacts, your files, your email and your browser history and adding this information to the storehouse of knowledge about you. Whenever you send or receive email, whether it is using Gmail or not, Google scans and records it, whenever you use a password to enter a site, conduct a banking transaction, use social media, buy anything at all or even go to porn sites, nothing is private, and Google keeps a record of your activities on its database, and nothing goes away.
And over time, and not very much time at all, Google can build up a very accurate picture of you, your life, your family, who you communicate with, what you do for work and leisure, how much you earn, how often you bank and most importantly, your behaviour as a consumer.
Google makes a massive profit selling this information to companies that want in turn to sell you goods and services, and they know so much about you and your behaviour as a consumer that it can tailor very specific advertising that matches your specific tastes, mood and income whilst you are searching the web, and even knows what triggers your buying behaviour.
Do you think it a coincidence that an ad for those just-my-style just-my-price jeans turned up on the news site you were browsing through?
No. Google is monitoring and refining its understanding of your consumer behaviour and what it knows about you from your online habits and your use of the internet over the years.
And it can’t be escaped, their software is so pervasive that even if you change computers, email addresses, your online identity, change browsers and even countries, it will eventually find you based on your known online behaviour and link your current information with its previous records about you, merge them and continue to monitor your activity.
Google knows you so well that according to its then CEO Eric Schmidt in an interview for the Wall Street Journal in 2007: “The goal is to enable Google users to be able to ask questions such as ‘What shall I do tomorrow?’ and ‘What job shall I take?”, and he went on to state that “I actually think most people don’t want Google to answer their questions, they want Google to tell them what they should be doing next.”
And that is a very scary place to be.
I want you to imagine someone who thinks they might have a serious problem with their digestive system typing into google “I have lower abdominal pain: what do I do next”. This ostensibly simple act has some pretty serious implications.
DIGITAL SOUNDS (0.05)
This example gives rise to the question, what has this got to do with healthcare?
Well, that amount of data can’t be ignored: in the right hands and interpreted correctly, it can tell us a whole lot about health, illness and disease.
We already know that these days, when a person begins to feel “off” their first port of call is the Google search bar, or Dr Google, as it seems to be called these days. Large studies suggest that for around 80% of people look to the internet and not their doctor as the first point of contact with the healthcare system when they feel unwell.
This means that companies like Microsoft and Google are starting to collate very large databases of search terms and key words that emerge from this activity. And remember, none of these search terms and keywords are stored in isolation: they all have associated pathways that run through the internet and go right back to the original computer they were entered into, and that means they are linked to you, and the database of your online behaviour.
So when you are ill, Google or Bing’s software flag subtle changes in your online behaviour well before you realise you are actually unwell yourself, and couple this up with keywords in your emails and social media posts and can actually start to analyse this at a deeper level and diagnose what’s wrong with you before you see a doctor.
Around five years ago, a number of independent researchers twigged onto the concept that a large part of the medical diagnostic procedure revolved around what a patient asks their doctor, and what words they use to describe what is wrong with them, and the doctor, drawing on their personal storehouse of knowledge and experience, matches the questions and keywords to a range of probable diseases that give rise to those questions in the first place and then begins to narrow the diagnosis down with more specific inquiries.
Now, they realised that these same questions and keywords were also being fed into google usually before a doctor was consulted, and if these questions and keywords were looked at in combination with changes in the online behaviour of an individual, clues began to emerge that could lead to a very accurate diagnosis.
Soon enough, Google, Bing and a team of independent medical researchers who had been given permission to ‘mine’ the data of both of these search engines began to develop diagnostic software that mines their massive database of search terms, keywords and general online behaviour to look for triggers that would flag a possible diagnostic event.
All without consent, mind you, since they actually own your online data: they just trolled though millions of gigabytes of internet history and the software looked for diagnostic clues and then narrowed its search down. When a diagnostic flag was raised, then the actual individual who was very probably unknowingly suffering from a disease could theoretically be traced and the person themselves identified.
So how successful was this project?
Well, in June of 2016, Microsoft announced that it had successfully and accurately diagnosed between 5 and 15% of all new cases of pancreatic adenocarcinoma in 2015 just by data mining the mass of internet records they owned for certain search terms, keywords and changes in online behaviour. It was even able to identify the individuals concerned and access their medical records (how, I don’t know) to confirm that diagnosis.
That is, just by trolling through the worldwide use of the internet, they were able to make an extremely accurate diagnosis of pancreatic adenocarcinoma in around 7960 individuals in the USA alone before those individuals had consulted a doctor or even knew they had the disease.
They were also able to look at the accuracy of their diagnosis confirmed that the incidence of false diagnoses was far, far lower than the average results for a physician.
Further, diagnosis, on average, was made at a point several months earlier than that which those developing the cancer would have sought ‘live’ medical advice.
Dr Google, so to speak, is now a reality, and your health is being monitored as we speak.
Where this starts to get really interesting though is when we look at the potential that lies within the process. A doctor makes a diagnosis by selecting the most likely possibility from the first lot of information given him by the patient, and then probes a little further to eliminate options and narrow the diagnosis down.
But Microsoft’s software can do this far more efficiently, accurately and far, far quicker without even asking a single question.
Without having to attend umpteen years of medical school, a software program can be loaded with a database of signs and symptoms, and then evolve, adapt and refine its diagnostic capabilities fifteen times faster than a consultant physician.
So for all ostensive purposes, Doctor Google is here, and it’s only a matter of time before diagnosis becomes an option offered along with google images and webs search bars, and it’s also inevitable that as deep learning takes hold, the scope of diagnosis and range of diseases that can be detected will begin to expand rapidly.
But, are there any issues we should take note of?
I think there are going to be problems and benefits with all of this, and in so many ways, diagnosis through mining internet data has a lot of issues that nurses need to be aware of.
There’s no doubt that this type of back end diagnostic tool is going to detect disease earlier and that will lead to earlier treatment ad better outcomes, especially since all the indications are that the software used is incredibly sensitive to the subtle early changes in user behaviour. But the downside of this is that if you aren’t a regular user of the internet, or don’t have access to the internet, you will be excluded from the diagnostic process. So the world will slowly divide into two camps: those who are regular internet users and those who are not, and we can expect higher morbidity and mortality amongst the have nots.
One of the other really big issues relates to patient confidentiality. Google and co own your browsing histories, and can examine data related to your health in their records without your consent, and without your knowledge, and this is not in violation of any privacy laws. This is a profoundly important point.
The reason this is an important issue is that there is a very high likelihood that companies like Google will form strategic (and profitable) alliances with specific healthcare providers, and on-sell data about you to them without your consent, much as they do now with your browsing histories and online behaviour for targeted marketing campaigns by retail and service providers. It’s entirely within the realm of possibility that as the diagnostic software detects early changes in your health, you’ll start to see ads appearing on your screen that subconsciously trigger you to buy healthcare products that are both within your budget, and address the very symptoms that are only just starting to niggle you now before some kind of notification pops up that suggests you visit your doctor. What I’m saying here is that there is a possibility that there will be a profitable strategy that determines how long after an illness or disease is detected that you should be notified.
This of course prompts the question: just who will google direct you to for treatment when it becomes apparent to their diagnostic software that something is wrong with you. As in the case of their advertising and ranked search results, will it become the case that the highest bidder gains prominence on your screen? Over time, that would present serious problems for an underfunded public health service.
Imagine, you have had a mild pain in your abdomen for a few weeks now, and very subtly, ads start appearing on your screen first for a specific brand of analgesia, and then for a medical service that specialises in lower abdominal pain.
One of the benefits of the internet that is associated with this issue is the increasing volume of information available to consumers. This means that they are better informed than ever before, and better able to make choices about their care. However, that increasingly presents some significant problems related to the unregulated nature of information available on the internet, and consumers are increasingly likely to present to hospital with pre-formed opinions and ideas about their treatment, irrespective of whether the information they use to form this opinion is of a reputable source or not. Nurses are going to end up spending an ever-increasing amount of time correcting opinions based on poor data, or worse, data that is profit driven and will need to draw patients back into the real world of limited resources and funding, the true determinants of treatment pathways.
Already IBM is claiming that their deep learning program can diagnose a wide variety of diseases with a far greater degree of accuracy and with fewer errors than an experienced physician. This is great, of course, but it foreshadows a significant reduction in the actual need for the role. Taken to extremes, it’s feasible that in the near future, the majority of medical decisions may be made electronically, without the need for human involvement. This has the potential to be an incredibly dehumanising experience for patients.
Finally, one of the biggest issues, in my opinion, is related to the shift of a large body of knowledge from a ‘live’ format, that is, in a person’s head and passed from generation to generation, to the electronic format. One has to consider what the outcomes would be if there was a war, a major disruption to the electricity grid, a major solar flare that renders electronic devices inoperative and if the internet goes down. That could potentially result in a complete collapse of the healthcare system, and if the data were not retrievable, then ultimately it would mean the loss of two thousand years of medical knowledge.
So doctor Google is here, already trolling through your browser history and looking at how you use the internet, searching for flags that could indicate you have the early stages of a whole range of diseases. This is going to be a huge part of healthcare in the near future, and nurses need to consider now how they are going to get involved, and how it is going to change their practice.
This is the Ausmed Handover podcast, my name is Darren Wake, and thank you for listening.
If you enjoyed this podcast, please subscribe to the channel for future episodes, and please feel free to leave your comments and feedback for us: we always welcome your opinion.