Ex Aula - Logo4-cropped


Nowadays, smartphones facilitate anything and everything, from sending emails to facial recognition.  But did you know these phones also have the capability to diagnose illnesses- sometimes even before the onset of any visible symptoms?

Neurological disorders affect hundreds of millions of lives each year. Although these diseases can largely be attributed to genetics, they are often caused by external factors such as malnutrition and parasitic infection.  This makes them extremely prevalent in lower-income countries where access to healthcare may be limited.  In addition to this, such conditions are currently expensive and complex to diagnose accurately.  This is where smartphones come in; the devices are increasingly being used to diagnose health conditions.  The large range of sensors they contain and the fact that so many people carry them makes smartphone-based monitoring an accurate, accessible and affordable solution.

The severity of certain neurological conditions, such as multiple sclerosis and Parkinson’s disease, is often related to a person’s ability to control their motion.  To measure how a person moves, we need some sort of reference point; the easiest way to do this is to look at how people walk- this is known as gait analysis.  Everyone has a similar walking style, and we can quite easily diagnose whether someone has a problem with their gait just by watching them walk.  If someone has a limping gait, for example, we can guess that they might have injured their leg. Gait analysis aims to be more specific about this, addressing questions such as: which leg is injured; which part of the leg is injured; how bad is the injury; how can we treat this.

Advancements in technology have made it possible to use motion-tracking computers to help us answer these questions. However, these computers are very expensive and only exist inside hospitals and research facilities. This means that current methods observe patients walking in time segments of a just few minutes (up and down the doctor’s surgery); this does not take into account factors that get worse over a longer walking period.  In particular, patients suffering from neurological disorders are often able to improve their gait when they concentrate purely on walking. To understand how we actually move, we need to find a way of monitoring these patients non-intrusively whilst they go about their day-to-day lives [1].

Anyone who has been affected by conditions such as multiple sclerosis and Parkinson’s disease will know that motion control slowly gets worse, often over many years with certain periods where symptoms can worsen or improve.  We have already established that measuring walking for short periods of time every couple of weeks or months doesn’t allow us to work out exactly how serious patients’ conditions are- and this only applies to those with access to healthcare.  So what’s the solution?

Smart watches and smartphones already provide us with a wealth of information about our health. They have the ability to measure heart rate, temperature, calories lost and much more.  Scientists in Oxford are now investigating whether or not it is possible to measure the long-term deterioration in the motor control of patients using data collected by a smartphone. To look at gait in more detail the smartphone component that we are most interested in is the accelerometer, which measures how ‘shaky’ the phone is. Buried deep within the millions of shakes that a smartphone experiences every day is a wealth of information on the gait of the patient and, by extension, their condition.

However, it is extremely difficult to analyse such large amounts of data, so we need to use machine learning. This involves teaching a computer what to look for in the data and lets the computer decide how severe a person’s gait abnormality is. Computers are far better than humans at handling this sort of data accurately, so a machine learning approach takes the workload off healthcare professionals.  Such a technique uses data sampled from patients and aims to develop an app which can be freely downloaded. The app will ask the patient to complete a couple of minutes of simple co-ordination tests and will then passively monitor them throughout the day. This will inform patients and their doctors about the progress of their neurological condition. For example, doctors will know how many times the patient fell over and how they handled different terrain, such as steps and obstacles. This will provide more accurate patient monitoring and result in a better quality of life regardless of demographic (patients can be released from hospital earlier if they are being monitored constantly at home).

Almost half of the global population carry a smartphone. This technology will be purely software based and will therefore come at no extra cost to both the users and the healthcare professionals. We are currently on the verge of a seismic shift in healthcare monitoring – gait analysis is just the beginning. This has untold benefits for society: doctors will be able to focus on the patients that need them most, while patients will be able to enjoy the freedom of their own home, only visiting hospital when absolutely necessary.

Of course, the use of smartphones in healthcare monitoring is not limited to neurological disorders; they are being recognised in the diagnosis of a whole range of conditions.  So, perhaps not everything our phones know about us is as detrimental as we think.

[1] L. Comber, R. Gavin, and S. Coote. Gait deficits in people with multiple sclerosis: A systematic review and meta-analysis. Gait and Posture, 51:25-35, 2017


– Harris Vince