Too Much Data In Medicine

I have recently started to feel overwhelmed with the amount of information being collected from me.

From my fitness tracker counting steps to my phone asking “How was your day?”, usually followed by graphs and messages indicating happiness, sleep quality and fitness goals I seem to always miss.

Although I signed up and invested in all these services, a question has started to form in my mind — is any of this data actually good for me?

I have dabbled with some personal data collection in the past (also know as the quantified self), and was until recently a very active participant even going as far to broadcast, collate and interpret my live heart-rate, location, Instagram feed and calories burned onto a website.

But how beneficial is knowing all this information really in terms of living a healthier and happier life?

What happens when we combine this data with healthcare?

Medicine And Your Data

In most parts of the UK and USA a typical visit to a doctor means you get a snapshot of your health at that particular moment. The process of making a diagnosis is based on the context of a patient’s medical history and readings taken on the day. This is used to make informed medical decisions.

With the exception of a few continuous monitoring devices and patient diaries, no infrastructure currently exists to cope with continuous data collection from environmental or personal health tracking devices which can impact on medical decision making.

Could Doctors Use My Data?

I remember (back when I worked as a doctor) patients coming into the A&E (ER) with pages and pages of spreadsheets full of data on everything from heart rate and symptoms to medications. Sometimes it was helpful, but most of the time trying to extract actionable information was really time consuming and not always in the patient’s best interests with regards to how they came into hospital in the first place e.g. a fractured toe.

Right now the average doctor doesn’t have the ability to make patient generated data mean anything.

Even if they did, the law is not clear on how they are allowed to use that information to treat patients.

If you use a device like a continuous ECG (electrocardiogram) to record heart rate, rhythm and character, and give that data to your doctor its not clear what they should do with it. They will have to work out if it should be part of your medical record, and if it does qualify it opens a whole host of other questions like:

  • Is the data compliant with local medical data standards?
  • How long does it need to be archived for?
  • Is it identifying information, or not identifying information?
  • Is it protected health information?
These questions require legal guidance, and sadly the law is always lagging behind technology in this area.

The Risk Of Over Medicalisation

Chances are if the average person got an MRI scan, a bunch of blood tests and a phycological evaluation they are going to find something wrong with themselves. Traditionally medics come from a mentality where the chief concern has been to guard against missing disease, with the focus on problems of under-diagnosis and under-treatment.

Once a pathology is identified, even in the context of an otherwise health individual — doing nothing can be viewed as negligence.

Right now this isn’t really a problem as only a small portion of population actively fund full body checkups. However, as health tracking technology becomes more readily available there is a risk that continuous data monitoring will create triggers resulting in more presentations to a healthcare service. Each time this happens, that person risks becoming ‘medicalised’, as they are subjected to testing and investigations.

In some cases this can be a good thing where such investigations may incidentally reveal a serious medical condition, but for most healthy people or people with known chronic conditions it may cause more problems than its worth and create unnecessary anxiety.

The health economist Alain Enthoven stated that increasing medical inputs will at some point become counterproductive and produce more harm than good. This is increasingly supported by mounting evidence about the threat to human health from over-diagnosis, the harms and waste from unnecessary tests and treatments, and data showing that the more a society spends on healthcare the more likely their inhabitants are to regard themselves as sick.

Our new data tracking toys, may be inadvertently creating opportunities to submit ourselves to institutionalised healthcare in the future.

Where Do We Go From Here?

Distinguishing the sick from the healthy has always been a challenge in medicine. With the inevitable increase in monitoring we must start to think about how useful this data will be in terms of informing clinical decision making.

Additionally we have to think about how this fits into the wider context of evidence based medicine (EMB) which balances clinical expertise, best external evidence and patient values and expectations to ensure the most appropriate and optimal decisions are being made.

Evidence Based Data

Looking at how medical treatment in the UK is evaluated as an example i.e. cost, effectiveness, patient population size and number of quality life years gained — I feel we could do something similar with data interpretation. Questions on data evaluation may be things like:

  • How common are brief arrhythmias are in the normal population?
  • How often our blood pressure might be high?
  • How widely normal oxygen saturations can vary?
  • The variation in the heart rate of an intrauterine baby?
  • What happens if a device/app malfunctions or is placed in the wrong position?
Understanding this information can help us make informed decisions about when something is safe or abnormal.

Regulation On Devices Or Apps With Diagnostic Impact

The US Food and Drug Administration and the UK Medicines and Healthcare Products Regulatory Agency have elected to go for light touch regulation on new apps and devices. We need clearer standards for what qualifies something as a reliable source of data when making diagnostic decisions.

Actionable Data Ranges

We need to utilise population and individual data clusters to give insights about the ranges of ‘normal’ within both a healthy and clinically diagnosed population.

Currently a business delivering any form of medical data interpretation will understandably have a low threshold for recommending that a user seek medical attention. Its within a business’s interest to do this to protect from litigation. By defining a ‘range of normal’ supported by governmental health institutions, we can improve the thresholds of ‘seeking medical assistance’, thereby reducing unnecessary patient anxiety and hospital admissions.

Final Thoughts

There is little point trying to resist the increasing data quantification of our daily lives. It won’t be long until these devices become embedded into everything from phones, beds, mirrors, clothes and shoes that data collection will happen autonomously.

The truth is that as of today these apps and devices are untested and unscientific, and will open the door of uncertainty.

Don’t be fooled: diagnostic uncertainty for most people can ignite extreme anxiety.

We must learn to work with data and ensure that decisions made about its interpretation are for genuine patient benefit. This not only includes the physical, but also the social and physiological aspects of medical care.

Like what you read? Give Adrian Raudaschl a round of applause.

From a quick cheer to a standing ovation, clap to show how much you enjoyed this story.