Tagged: small data

How Edward Snowden Might Impact Big & Small Data Use in Healthcare

In the two weeks before Edward Snowden broke news about PRISM I gave two lectures about research ethics. It is traditional to note that what is ethically acceptable is in constant flux as technology changes and also that ethical principles tend to emerge in times of crisis, e.g. human experimentation that is also genocide. An observation that I made during these lectures was that the next crisis is probably going to relate to the use of ‘big data’ repositories in health. Well I didn’t quite nail the health part but bingo! – the use, security and privacy of individual’s data, in this case meta-data from telecommunications, email and text messaging, became the topic of the day.

The public is now aware that there are large repositories of data about themselves. There will start to be greater awareness that some of these data repositories include health records that are collated electronically. Currently, in Australia, these repositories and most other healthcare systems, these repositories are not very detailed but they do record if you’ve been to hospital, why you’ve been to hospital and a host of demographic information. This data is collected legally and is used legally for public health and planning purposes and in many cases for medical research, usually under additional regulation through ethics review processes.

The other bit of information that the public will become aware of is that these repositories of data are linkable. This means that a dataset collected in one system like healthcare is potentially linkable to a dataset in another sector like finance, or the justice system.

The issue becomes how the public, how society reacts to this knowledge and what it believes is reasonable use. The argument in the op-ed pieces in the major publications is that programs like PRISM serve a purpose in protecting the public from security threats. Indeed, surveys of the public in the US suggest that generally speaking people support this assertion. So there are some ‘reasonable uses’ in the public interest. No doubt, the same will apply to health data. The question is where the thresholds lie and how much transparency there will be.

But let’s get back to all of that metadata. Clearly there might be uses for tracking terrorists or gun-owners. But there might be healthcare applications. The concept of small data has recently emerged. The example is the notion that your phone company can track your movements through your mobile phone. If an older person was less active this could show up in their small data and indicate that they are sick. Who does this data belong to? The person or the phone company?

At the moment there are no answers but no doubt there will be trickle down implications from the Snowden affair to other applications like healthcare.

Towards the minimally disrupted quantified self

One of the emerging movements in healthcare is the quantified self movements. Led by a bunch of tech-savvy self-confessed geeks the quantified self movement is all about collecting personal data and using it to inform life decisions, health care decisions and behaviour change to improve overall well-being. These life-loggers are measuring their habits, food intake, activity levels, mood, heart-rate, blood pressure, blood sugar and a host of other person-reported outcomes and physiological measures and charting them on paper or in spread sheets. Whilst this seems like a small sub-culture most people with smart phones have downloaded at least one health related app that can be used for life-logging. Persistent use is not always durable and perhaps it is because life-logging, like much other healthcare activity is burdensome. You have to spend a lot of time entering data yourself. Well maybe this is all starting to change and being a quantified self is becoming less burdensome.

Victor Montori has been advocating for minimally disruptive medicine, the aim being to reduce the overall burden of healthcare (not just the burden of disease) for patients, particularly those with chronic medical conditions like diabetes. Many chronic medical conditions actually already require individuals to include some form of life-logging as part of their healthcare routines e.g. recording blood sugar levels or blood pressure.

Smartphones and other smartphone facilitated devices are starting to enable data collection in a minimally disruptive way. For example accelerator and GPS-enabled smartphones, or smartphones linked wirelessly to other small devices like a Fitbit or a Nike Fuelband, can track exercise and other activity levels. Calorie consumption can be inferred and flights of stair counter. The same phone or added devices can be used to measure sleep patterns. Increasingly attachments to the phones can measure various physiological parameters in a non-invasive way. Apart from the occasional re-charge (they aren’t solar or movement powered yet) they are pretty much plug & play, set & forget. All you need to do is make an effort to check out the data in the form of customised reports and decide how you will act on them.

To take it a step further companies such as Soma Analytics have apps that move beyond you recording your mood directly to actually inferring your mood by analyzing your voice patterns during phone conversations. This in turn allows a calculation of your stress levels and can give you feedback and advice about stress management.

Another variation is to use the phone using the GPS tracking facility to record activity levels as a marker of a patients’ “social pulse”. For example, if an older person changes their activity patterns based on monitoring of the phone movements then this might be a sign of decline in functional status. Deborah Speaking at @TEDMED Deborah Estrin has referred to this as our “small data” and is encouraging processes to obtain this data from telecommunications providers.

All of this data can be used for personal use but uploaded to the cloud and given the right permission your medical practitioner could access the data remotely and in return provide analysis, interpretation and clinical advice on the basis of your data. Automated alerts flowing to healthcare provider or patients could be a form of minimally present telemetry.

Looking at these existing and emerging technologies one can see that having a data-driven life doesn’t need to be a full-time job even if you are collecting data 24/7/365. Even the activities which might require an effort at data entry, like recording diet or medication use, might become more automated through image recognition software or barcode/QR-code scanning. This could be continuous and uninterrupted for individuals using google glass. I think we can call this the minimally disrupted quantified self.

Minimally Disruptive Medicine

Quantified Self

Small Data

Soma Analytics